Instructional planning using focus of attention

Abstract

An instructional planner needs to revise both the student model and the instructional plan itself as its perceptions of the student change during the interaction with the student. Artificial intelligence provides a variety of reason maintenance systems (RMSs) whose job is to carry out such revisions. Unfortunately, traditional RMSs cannot be used directly in real time instructional planning because they are typically quite slow. To overcome this problem, we propose a new RMS, called the attention-shifting belief revision system (ABRS), that works efficiently by focusing only on the parts of the student model and the instructional plan that are relevant to the current subgoal(s) of the planner. The planner identifies current goals and relevant beliefs in the student model. The ABRS then ensures that these goals and beliefs are kept in focus and that their consistency is maintained. An example shows that instructional planning using the ABRS is just as effective as using traditional RMSs, but is considerably more efficient.