A Framework for Model-Based Adaptive Training

The Seven Modelling Dimensions

Qualitative Modelling requires the definition of an explicit model of the physical device, product or process on which application tasks are to be performed. In a training application this model-based approach allows flexible learning, including unforeseen and therefore unspecified situations.

An adaptive model-based approach enables a training system to be applicable for a wide range of training tasks. The specified expertise can be demonstrated and validated with respect to model(s) rather than on ‘compiled’ pre-specified expertise focused on particular training tasks.

The internal (computer-based) model of the training environment can be considered as either a complete domain model or set of case models. A domain model is a set of objects that represent the training domain.

As in Proto-MOBAT, for Workmanship-MOBAT the domain model is easier represented as a set of examples or a set of cases (which together make up the domain model). A case model is a particular situation being reasoned about in the domain. The technique for creating a set of case models is easier for the workmanship domain because the domain model can evolve and gradually be extended as needed.

A set of primitive model properties from the characterisation of multiple models in ITSIE is adapted below. The ‘model’ is an executable representation of training subject matter. Depending on the subject matter, various concepts can be used to describe the primitive model properties.

Each of the seven modelling dimensions is mapped to aspects of the Workmanship-MOBAT training application which is discussed in separate sections.

Ontology refers to the source of knowledge. I.e., model descriptions consisting of concepts with attributes, relations and expressions.
Scope refers to the part of the domain within which training takes place.
Generality refers to how broadly the model can be applied.
Perspicuity refers to the ease-of-use of the model representation.
Precision refers to the exactness of the detail in the model description.
Accuracy refers to how near the model description is to a reference model.
Uncertainty refers to the belief in the model description.