Intelligent Tutoring Systems,

Intelligent Tutoring Systems, Tutoring Module,Expert Module

—Intelligent Tutoring Systems are systems which pro-vide direct customized instruction to students. An IntelligentTutoring System consists of four modules. This research concen-trates on two of the modules, namely, the Tutoring Module andthe Expert Module, analyzes tutoring systems which contributeto the design of these modules. The study identifies the issueswhich have not been addressed in previous contributions.

 Index Terms

—Intelligent Tutoring Systems, Tutoring Module,Expert Module

I. INTRODUCTION

An intelligent tutoring system (ITS) is any computer systemthat provides direct customized instruction or feedback tostudents, i.e. without the intervention of human beings, whilstperforming a task. ITS typically consist of four modules [5]:the Expert Module, which comprises of facts and rules of theparticular domain to be conveyed to the student; the TutoringModule, which designs and regulates instructional interactionswith the students; the Student Module, which is a dynamicrepresentation of the students current state of knowledge; andthe User Interface, which controls interaction between thestudent and the system. The design of an ITS can focus onvarious issues, including the tutoring decisions which takeplace in the tutoring module and the facts and rules representedin the expert module. The following study discusses each of these modules as well as tutoring systems which address thedesign of these modules, as well as identifies the issues whichhave not been addressed in previous design of these modules.The following figure illustrated the structure of an ITS.

II. TUTORINGMODULE

The tutoring module is the instructional module that designsand regulates instructional interactions with the students. It isclosely linked to the student model, using knowledge aboutthe student, and its own tu torial goal structure to decidewhich pedagogic activities will be presented: hints to over-come impasses in performance, advice, support, explanations,different practice tasks, tests to confirm hypotheses in thestudents model, etc. Assessment is a very important functionof the tutoring module. The function of the tutoring module isessentially to perform continuous assessment of the student,and thereby interact with the expert module to prescribe furtheraction.Decisions on further instruction are taken in the Tutoring Mod-ule. The decisions are taken based on various characteristics.Some tutoring systems take decisions based on the studentsperformance and teaching history [1], while others decidebased on students emotions, known as Affective Modeling [2].Learning habits and learning targets are also taken into account[3].Example-tracing tutors make use of a ”behaviour graph” inthe tutoring module, which is a directed, acyclic graph thatrepresents multiple ways of solving the problem [4]. Basedon the method chosen by the student, decisions are taken.The categories of Intelligent Tutors - Curriculum Sequencingand Tutoring Strategies identified by Tom Murray also addressthe Tutoring Module. Systems which belong to the categoryof Curriculum Sequencing consists of rules, constraints andstrategies for sequencing courses. Sequencing and rules arenot necessarily decided by the author, and can be decideddynamically. Systems which belong to the category of TutoringSystems incorporate micro-level tutoring strategies, sophisti-cated sets of instructional primitives, and multiple tutoringstrategies. Granularity of hints and explantions is small. Tu-toring rules are decided by the author. Domain knowledge of systems belonging to both categories are typically shallow.COCA is also an intelligent tutoring system [7] which placesimportance on both authoring of the domain as well as how thematerial should be taught. In its Tutoring Module, COCA alsoattempts to simulate the decisions a teacher might make duringthe tutorial process, such as a decision point about the natureof the student’s next activity (instruction or assessment),