Affiliations: Intelligent Systems Technology, Inc., Los Angeles, CA, USA
Abstract: Intelligent agents are increasingly being exploited as synthetic
role players (e.g., teammates, coaches, and opponents) in scenario-based
training (SBT). However, the introduction of intelligent agents into training
applications comes with its share of challenges given the unique requirements
of each role. In particular, the required level of intelligence depends on the
requirements of a particular role. For example, agents performing as
"extras" in a scene require little or no intelligence,
whereas an in situ remediation agent requires the ability to diagnose student
deficiencies relative to, for example, the school solution and provide
appropriate instruction to remedy the deficiencies. To perform some functions,
agents require an understanding of the problem domain; for others, they merely
need to understand their own tasks. In some cases, agents need the ability to
plan, whereas in other cases they merely need the ability to react. A
pedagogical agent that typically interacts with the learner stands to benefit
from having a human persona (an avatar) and an instructor-like personality. In
light of the foregoing, the implementation of these agents for scenario-based
training (SBT) can take a variety of forms and vary in their level of
sophistication. This paper presents the various roles that intelligent agents
can play in SBT to dramatically lower training costs without compromising
training effectiveness.
