Physics-Based Human-in-the-Loop Machine Learning Combined with Genetic Algorithm Search for Multicriteria Optimization: Electrochemical CO2 Reduction Reaction

Abstract

Machine learning (ML) can be a powerful tool to expedite materials research, but the deployment for experimental research is often hindered by data scarcity and model uncertainty. An human-in-the-loop procedure to tailor the implementation of ML for multicriteria optimization is described. The effectiveness of this procedure in the development of a nafion-based membrane electrode assembly for electrochemical CO2 reduction reaction (CO2RR) into CO for two targets is demonstrated: energy efficiency (EE) and partial current density for CO2RR ((Formula presented.)). Model-agnostic nonlinear correlation analyses identify the 11 features relevant to those targets. The three studied decision tree-based ML models yield similar cross-validation errors so an ad hoc feature analysis of the models is done with SHapley Additive exPlanations and nonlinear correlation techniques. The predicted EE- (Formula presented.) space and the functional dependency of the predictions are investigated to assess model plausibility. A genetic algorithm with CO production cost as the final target with subsequent validation experiments of candidate conditions is devised. The model chosen through ad hoc analysis yields the highest accuracy and the only one that can locate the Pareto front with a single round of experiments, demonstrating how appropriate model selection through careful inspection can greatly accelerate the research cycle.