Convection-diffusion-reaction EKI and MF-EKI ============================================ **Method.** Ensemble Kalman inversion (EKI) estimates unknown parameters from observations using an ensemble of model evaluations. Multifidelity EKI adds cheaper surrogate evaluations. **Example.** The example infers diffusion and reaction parameters in a steady convection-diffusion-reaction model, comparing full-order EKI with projection-ROM, Gaussian-process, and neural-network multifidelity variants. Run: ``python examples/eki_mf_eki_demo/example.py``. .. figure:: notebooks/eki_mf_eki_demo.svg :alt: EKI and MF-EKI convergence on the CDR problem :width: 90% Observation error across EKI variants. For algorithm details, settings, and full source listings, see :doc:`eki_mf_eki_demo_details`. .. toctree:: :hidden: eki_mf_eki_demo_details