"Tutoring systems are described as having two loops. The outer loop executes once for each task, where a task usually consists of solving a complex, multi-step problem. The inner loop executes once for each step taken by the student in the solution of a task. The inner loop can give feedback and hints on each step. The inner loop can also assess the student's evolving competence and update a student model, which is used by the outer loop to select a next task that is appropriate for the student. For those who know little about tutoring systems, this description is meant as a demystifying introduction. For tutoring system experts, this description illustrates that although tutoring systems differ widely in their task domains, user interfaces, software structures, knowledge bases, etc., their behaviors are in fact quite similar."
Kurt Vanlehn. 2006. The Behavior of Tutoring Systems. Int. J. Artif. Intell. Ed. 16, 3 (August 2006), 227-265.
1. SUMMARY
The paper is about behavior of intelligent tutoring system (ITS), what such systems tend to do, and some important design options. The paper describes 6 tutoring systems with chronical tasks. Most system require students to perform steps in order to solve tasks. The paper then raised the concept of outer loops (dealing with tasks) and inner loops (dealing with steps). Systems post tasks (outer loop) and give hints/feedbacks in the inner loops.
There are four common types of outer loops: menu based (selected by learners), predetermined sequence (tutor assigned, fixed), mastery learning (tutor assigned, flexible), and macroadaptive learning (tutor assigned, adaptive to learning traits). The problems include (1) providing adequate task bank (2) designing effective algorithms to select tasks.
Common services within inner loops are: minimal feedback for steps, error-specific feedback, hints, assessment of knowledge, review of the solution. The step analyzer and the step generator drive those services. Minimal feedback categories include: immediate feedback, delayed feedback, demand feedback. Hints should only be given when learners really needed or specifically requested. Hints should honor instructor’s preference. Error specific feedback can be error descriptions, or a sequence of hints. Assessment of knowledge can be coarse-grained for fine-grained (by counting success learning events and counting failures).
The paper identified that most tutoring systems do not have the inner loop. A single loop system might be a good choice in terms of system development time v.s the amount of student learning per unit. While the tutoring systems are different, their behaviors are quite similar.
2. STRENGTHS
The paper is very detailed, having a lot of important citations, providing great backgrounds on tutoring systems. The paper is well-written and easy to read. Main ideas can be spotted quickly.
3. WEAKNESSES
The abstract could have been more detailed, more developed to match with the complexity of the topics mentioned.
Image credit: The Behavioral Scientist
