A Framework for Model-Based Adaptive Training

Personal Details - Trainee Profile

The trainee profile is basically a record of a trainee’s knowledge and a static database relating the trainee to an employee within the industrial environment. In the MOBAT specification framework the main trainee profile categories are discussed below. These categories are mapped to system internal slots in the trainee model for use by the trainer and expert training agents. The training system may need to be able to identify individual trainees. Recording and recognising individual trainees is useful if the trainee is using the training system for future training sessions. Individual trainee specific information may include such data as: first name, last name, badge number or employee identification number, photograph of the trainee, age, sex and date of birth. The training system may need to establish how relevant the training subject is to a trainee’s job. Depending on the training application area, the trainee job area and job category can be chosen from domain dependent values. For the trainer agent these values can be mapped to generic job relevance information such as (a) high, (b) medium and (c) low.

Computer based training systems can use various categories to specify a trainee’s experience and ability in performing tasks in the subject domain. Although a more detailed mapping to determine a trainee’s self-supporting level is possible in the MOBAT specification framework, in general a trainee’s experience in the subject area can be chosen from categories such as: (a) beginner – a novice who has not finished the theoretical training; (b) competent – a trainee with a medium level of understanding who has completed the theoretical training but does not have sufficient practical experience; and, (c) advanced – the trainee has experience in the training subject area and can interact at a more complex level with the expert agent. These categories are described further below for application in the MOBAT framework.

• A beginner trainee is a novice in the domain who has not finished the theoretical training. The trainer agent (didactic tactician) may suggest an introduction to the training tasks with presentation training units. The expert agent needs to be prepared to answer trainee questions and to provide detailed explanations. A novice trainee may need to be given an introduction to the training subject, therefore the trainer agent (didactic tactician) may select an appropriate model and focus the problem arena for the trainee (i.e., select the scope modelling dimension). A novice trainee may have misconceptions about the training subject, therefore the trainer agent (diagnostic tactician) may need to remediate misconceptions and provide alternative portrayals of presentations (i.e., adjust the uncertainty modelling dimension).
• A trainee can be considered competent in a domain with a medium level of understanding. This trainee has completed the theoretical training but still does not have sufficient practical experience. The trainer agent (didactic tactician) may provide example and practice training units. The expert agent is required to provide advise and supervision to help the trainee practice training tasks. A competent trainee may need to be capable of using task expertise in different situations, therefore the trainer agent (didactic tactician) may suggests the use of appropriate case-models for the training task (i.e., select the perspicuity modelling dimension). A competent trainee may consistently have a particular problem (i.e., a ‘bug’ can be classified), therefore the trainer agent (diagnostic tactician) may intervene and provide the trainee with a new problem range (i.e., adjust the accuracy modelling dimension).
• An advanced trainee has experience in the training subject area and can interact about training problems with the expert agent. The trainer agent (didactic tactician) may make frequent use of question training units. An advanced trainee may need to be capable of performing multiple ways to do a task, therefore the trainer agent
• didactic tactician) may switch models to assess the trainee using different problem solving techniques (i.e., select the generality modelling dimension). An advanced trainee may occasionally make errors (i.e., a ‘slip’ can be detected), therefore the trainer agent (diagnostic tactician) may test a trainee’s understanding at another level in the domain model description (i.e., adjust the precision modelling dimension).

The trainee’s interest in the training subject matter indicates the trainee’s level of motivation to learn about the subject. This can be chosen from values such as: (a) high interest in subject; (b) moderate interest in subject; (c) low interest in subject; and, (d) casual interest in subject.

The trainee’s familiarity with, or confidence in, using the training system interface can be expressed in categories such as: (a) low – not familiar with training system interface; (b) medium – have used the training system before; and, (c) high – knows how to use the training system interface.

A trainee’s learning ability can be chosen from values such as: (a) average learning ability; (b) fast learner; and, (c) slow learner. A start has been made in the training framework to define and map all of these categories into a trainee self-supporting level but it is by no means complete, there is still much research to be done.