PyMED: Multiphase Materials Exploration via Evolutionary Equation Discovery

M²E³D discovers equations, models and correlations underpinning experimental data. It builds on the fantastic PySR and fvGP libraries to offer a user-facing package for:

• Discovering symbolic closed-form equations that model multiple responses.
• Efficient parameter sampling for planning experimental / simulational campaigns.
• System multi-response uncertainty quantification - and specifically targeting high-variance parameter regions.
• Automatic parallelisation of complex user simulation scripts on OS Processes and distributed supercomputers.
• Interactive plotting of responses, uncertainties, discovered model outputs.
• Language-agnostic saving of results found.

M²E³D was developed to discover physical laws and correlations in chemical engineering, but it is data-agnostic - and works with both simulated and experimental results in any domain.