Electrochemical experiments produce rich, high-dimensional datasets — impedance spectra, cyclic voltammo- grams, galvanostatic cycling profiles, chronoamperometric transients — that are difficult to interpret fully by conventional analytical methods. Machine learning offers new routes to automated feature extraction, degra- dation diagnostics, state-of-health estimation, and closed-loop protocol optimisation, with the potential to accelerate both fundamental understanding and practical deployment of electrochemical technologies.

This workshop examines the state of the art in ML-assisted electrochemical analysis. It addresses the landmark studies on data-driven prediction of battery cycle life from early-cycle data, Gaussian process-based analysis of electrochemical impedance spectroscopy for degradation mode identification, and Bayesian closed-loop optimi- sation of fast-charging protocols — studies that demonstrate the power of ML to extract information that is present in electrochemical data but inaccessible to conventional analysis.

The session also explores ML applications for molten salt systems — conductivity, viscosity, and phase behaviour prediction — and discusses the challenges of open datasets, reproducibility, and standardised benchmarks in electrochemical ML. Participants will develop practical knowledge of how to apply ML methods to their own electrochemical data, an understanding of the validation requirements for credible results, and an appreciation of the opportunities for ML to transform electrochemical science and engineering.

Extended Key Takeaways & Outputs

• Practical strategies for ML feature extraction from cyclic voltammetry, EIS, and cycling data

• Understanding of landmark results in battery lifetime prediction and protocol optimisation

• Knowledge of GP-based and Bayesian approaches to electrochemical diagnostics

• Appreciation of ML opportunities for molten salt and high-temperature electrochemical systems

• Guidance on open datasets, reproducibility standards, and validation for electrochemical ML