v1.0.0 — Manuscript submission snapshot

Archival snapshot of the code accompanying the manuscript:

"Quantum reservoir computing for battery state-of-health estimation from electrochemical impedance spectroscopy" (submitted to [TBA]).

This release freezes the full analysis pipeline used to generate every figure, table, and statistical result reported in the manuscript and supplementary material.

What this repository contains

• Data loaders for three EIS + capacity datasets: Stanford SECL (public), Warwick DIB (public), and an ESCL laboratory dataset (restricted; metadata only).
• Quantum reservoir computing (QRC) pipeline: arctan feature encoding, CZ-ring entangling circuits, and statevector / shot-based expectation-value readouts (Z, Z+ZZ, XYZ observables) over 6 qubits.
• Classical baselines: Ridge, SVR, XGBoost, Echo State Network, Random Fourier Features, and PC1 linear regression, evaluated under identical LOCO and temporal protocols.
• Noise-model simulations (Qiskit Aer, IBM Marrakesh depolarizing + readout error model) and real-hardware validation on IBM Quantum devices via QiskitRuntimeService + SamplerV2.
• Statistical supplement: Wilcoxon signed-rank tests, bootstrap confidence intervals, Cohen's d, nested LOCO cross-validation, and kernel-target-alignment / effective-rank / t-SNE feature-space diagnostics.
• Phase runners (src/phase_1 through src/phase_12) reproducing every analysis end-to-end from the raw .mat files.

How to reproduce

conda activate escl-quantum
python -m src.phase_1.run_phase_1
python -m src.phase_4.run_phase_4 --noiseless-only
# ... see README.md for the full phase list

Each runner writes logs to result/phase_N/data/log.txt and figures (PNG + PDF) to result/phase_N/plot/.

Citation

If you use this code, please cite the manuscript (citation to be updated upon acceptance) and this Zenodo archive.

License

Released under the Apache License 2.0 License — see LICENSE.

Acknowledgements

IBM Quantum services; the Energy Storage and Conversion Laboratory (ESCL) at Chungnam National University; and the curators of the Stanford SECL and Warwick DIB public datasets.