A Framework for Model-Based Adaptive Training

Preferred Trainee Learning Style and Learning Mode

The preferred trainee learning style and learning modes are considered a separate part of a trainee profile. Each aspect is described in this section.

The trainee learning style is a record of the preferred way of learning and interacting with the training agents. A trainee’s preferred learning style can be specified as a classification of styles such as: learning from doing a task, learning from examples, learning from presentations, learning from problem solving, learning from analogy and learning from memorisation.

This learning style classification can be useful to provide a way of specifying trainee preferences with readily understood descriptions. A trainee’s learning style may include preferred ways of communication with a set of context sensitive help tools. The MOBAT specification framework is based on an agent architecture where different training agent as well as communication channels to other users may be available to the trainee.

The context sensitive help agents are called “guides” in [MOBIT 1994, W1, W6]. The communication style is chosen from: (a) a guide for personal advise (b) a roundtable of available guides; and, (c) a roundtable of other users. The use of these ‘guides’ within the training system architecture (see Figure 1-1) has not yet been fully established. Initial results for the use of ‘guides’ are presented in [MOBIT 1994-1996].

The specification of models at a level appropriate to a trainee’s goal expertise level provides a view of learning in three categories, which are called the preferred learning modes. The learning mode is based on the mapping between available knowledge and goal expertise. The knowledge categories are:

  • Procedural Knowledge (K1);
  • Associative Knowledge (K2); and
  • Principled Knowledge (K3).

The goal expertise categories are:

  • Skill-based Expertise (E1);
  • Rule-based Expertise (E2); and
  • Model-based Expertise (E3).

The learning modes are determined from the mapping Knowledge and Expertise pair sets as shown in Figure 3-2 Types of Learning Styles (Learning Modes).

The learning mode specification has implications for the development of both the expert and trainer agents. If the representation of knowledge at different levels is required (and it is possible to acquire the knowledge to do this), then the appropriate tools for domain expert representation together with the selection of learning modes provides a rich learning environment based on the evolution of models as training progresses.

It should be noted that even with one learning mode the trainer agent can use several model varying techniques for achieving a particular level of performance. These three learning modes promote different types of learning. For a further description of these learning modes see Section 3.3.2 [ITSIE 1992, D7].

Flower Show
© | | Sitemap