A Framework for Model-Based Adaptive Training

Training System Terminology

There are many components that make up a specification framework for Intelligent Training Systems. As many isolated groups have researched ITS technology, for different purposes, and in different cultures, there are many sets of terms in use. Also, within the MOBIT project different terms have evolved to describe the same components. The terms below will be used in this report. This is not an attempt to define the “correct” and complete set of terms, but merely a set that is consistent and that can be used as a starting point to describe the MOBAT specification framework.

Available Knowledge . Knowledge which, if possessed by an agent, allows the fulfillment of domain tasks. Available knowledge is represented with domain models within the Expert Agent of a training system.

Domain Model. A representation of the training subject matter that is constructed for the purpose of solving a particular problem, or set of problems. Domain abstraction levels are chosen from procedural, associative and principled model types.

Expert Agent . The software component of an intelligent training system which embodies (in executable models) the domain specific available knowledge required for training.

Goal Expertise . The expertise sufficient for a trainee to achieve the training objectives.

Knowledge Role. The specification of the purpose of a chunk of knowledge within a problem solving method (see Section 2.5.2 Role Limiting Methods).

Learning Mode . Represents a mode of achieving trainee goal expertise by selecting from the subject matter the appropriate available knowledge (chosen from deductive, inductive and rote learning modes).

Problem Space. The subject arena within which a trainee is expected to achieve a certain level of performance (see Section 2.3.1 Cognitive Modelling in Problem Spaces).

Problem Solving Methods . Methods used by the expert agent or the trainee which control the way to achieve a task in the problem space (see Section 2.5.3).

Skeleton Training Plans . Procedural plans which are provided to the trainer agent via the trainer interface.

Trainee Interface . The user interface for the trainee.

Trainee Model . A model which reflects a trainee profile (e.g., knowledge, ability, motivation and confidence) and trainee preferred learning modes which allows a trainer agent to act.

Trainer Interface . The user interface for the human trainer for skeleton training plans.

Trainer agent . The software component of an intelligent training system which selects, directs and monitors the execution of training units so as to achieve the training goals.

Trainer Tasks. The actions used by the trainer agent to implement skeleton training plans and training units.

Trainer Methods . Methods used by the trainer agent which control the way to execute trainer tasks.

Training Agents . The invariant structure of trainer and expert agents, relating training objectives, trainee model and domain models via a set of cognitive processes.

Training Goals . Goals expressed in terms of representations of the training objectives.

Training Objectives . Informal descriptions of the behavioural capabilities which are to be expected from a trained individual.

Training Strategy . The style of interaction adopted by the trainer agent which may be realised by trainer methods and trainer tasks (chosen from tutoring, coaching, facilitating and discovery training strategies).

Training Units . Objects implementing the training dialogues of the trainer agent.

Transfer of Learning. Learning methods embodied within the training system itself (chosen from within-trial, across-trial and across-task – see Section 2.6.1).

Flower Show
© | | Sitemap