Convex Hybrid Modeling: An Operator-Based Approach
Researchers have developed a new approach to hybrid modeling that combines the accuracy of machine learning with the interpretability required for decision-making systems. This method formulates convex learning problems to systematically account for interpretability, offering efficient surrogate models. The approach utilizes operator theory to re-parameterize models in a "lifted" space, treating the system as a kernel-based mixture of interpretable models, with applications demonstrated in both static and dynamic models. AI
IMPACT Introduces a method to create more interpretable machine learning models for decision-making applications.