How to Create Admissible Set of Functions?

Creating an admissible set of functions is a fundamental challenge in learning theory, as it requires a small VC dimension and clever selection methods. Traditional statistical learning theory often assumes that these sets are predetermined; however, this approach overlooks the complexity involved in deriving such sets from a vast continuum of functions. Success in this domain necessitates an understanding of the properties of training data and the relationship between the models and their expected values. The issue revolves around identifying appropriate functions that accurately reflect the training data, necessitating the formulation of specific questions that guide the inquiry. The process also involves recognizing general patterns related to specific problems, which is central to developing intelligence in machine learning systems.