Traditional ensemble methods such as Mixture of Experts, Bagging, Stacking, and Boosting can be reinterpreted within the Functor Model Architecture (FcMA) as structured compositions of morphisms. In this formulation, ensemble behavior is not the result of repeated training over parameterized models, but arises from compositional algebra over executable mappings. The familiar function-level equations are retained as concrete realizations, while the broader interpretation lifts these constructions into functorial mappings over categories of inputs, contexts, and executable artifacts. This perspective clarifies compositional semantics, improves auditability, and enables localized updates through controlled morphism application.
John Harby (Sat,) studied this question.