JAIED v8n1

The purpose of academic advising is to help a student develop educational plans. In universities it is usual for faculty to perform academic advising. Recently a number of rule-based systems were developed to help faculty and students with the advising task. Some of these systems were reviewed in our previous paper, where we posed a number of major goals for constructing knowledge-based advisors. SAIKIC, our academic advising system, is designed to incorporate those goals. In this paper we describe our progress with the design of SAIKIC.

Return to Contents

An Expert Advisor for Vocational Guidance

JOACHIM P. HASEBROOK

Bank Academy, Oeder Weg 16-18, D-60318

Frankfurt, Germany

WOLFGANG NATHUSIUS

Medialog Corporation, Kolpingstr. 18, D-68165

Mannheim, Germany

We developed a multimedia program which combines a vocational encyclopedia and a testing facility to foster adequate career decisions. The testing facility is designed to suggest the same careers which a given number of experts would have suggested if presented with the same user´s input. Our vocational database includes imprecise data, like expert ratings, enabling the calculation of suggestions of career options. The most important group of software users for vocational guidance are young adults who are about to leave school. The results of cluster analyses (n=426) showed that the interests of students were poorly structured and were not compatible with experts´ ratings. The test facility has been implemented on several CD-ROMs, in a short quiz to identify occupational fields, and in a wide range of surveys which was answered through letters. Forty-three students participated in an experiment to investigate the understanding and acceptance of the information provided by the system. The results showed that students were able to judge whether careers matched their individual interests or not. Furthermore, we explored whether the system was able to reconstruct 38 experts´ ratings. The system showed a good performance in reconstructing the experts´ data&emdash;except with one academic career which was not described very clearly. In a recent study, we tested the influence of the testing facility on recall of information and individual acceptance (n=75). Acceptance and recall about career options were clearly enhanced, when studying individualized materials compared to general information.

Return to Contents

Toward Computational Models of Motivation: A Much Needed Foundation for Social Sciences and Education

MICHEL AUBÉ

Université de Sherbrooke

Sherbrooke, Quebec J1K 2R1, Canada

Most basic concepts about knowledge representation, such as declarative or procedural knowledge, come from research in artificial intelligence and have percolated from there through other cognitive sciences to finally reach the domain of educational practice. Therein they are beginning to have a profound impact upon curriculum design, didactics, teaching strategies, and teacher education. Meanwhile, the domain of student motivation is calling for increasing attention, but existing models remain essentially descriptive: They relate the variables that are known to matter empirically, yet they offer little explanatory power about what activates willful behavior and goal orientation. In this paper we advocate the need for basic computational models of motivation as new fundamental tools for education and suggests using the design-based approach of artificial intelligence to lay down such models. Motivation includes physiological needs as well as emotional reactions, but our proposal for a model will only be concerned with emotions. It is argued that motivation has to do with resource management, and that the particular resources that concern emotions are the ones that are obtained from others, through their having been committed to provide them. Thus viewed, emotions are essentially of a social nature and have to do with regulating and managing commitments that bind individuals together in action, communication, and cooperation. Preliminary specifications for a computational model of emotions are formulated, and consequences for current theories of emotions and of education are formulated in conclusion.

Return to Contents

Analogy, Logic Programming, and Metacognition

ANTONIO M. LOPEZ, JR.

Mathematics and Computer Science

Loyola University, Campus Box 51

New Orleans, LA 70118-6195, USA

This paper presents an analogy microworld used in an artificial intelligence (AI) course to stimulate student metacognition. Students write programs in a logic programming environment that can answer queries about analogies in the microworld. In developing these programs, the students prepare interim reports that explain their thinking. With these reports the instructor can guide and clarify the student’s thinking. In an AI course, the cognitive effects of programming computers to do intelligent things greatly depend on the mindful engagement of the students on these tasks.

Return to Contents

Using Students’ Knowledge to Generate Individual Feedback: Concept for an Intelligent Educational System on Logistics

DIETRICH ZIEMS AND GABY NEUMANN

Department of Logistics and Materials Handling Engineering

Otto-von-Guericke-University of Magdeburg

Postfach 4120, D-39016 Magdeburg, Germany

Engineering education mainly focuses on enabling the student to deal independently with complex and complicated problem-solving processes in a varying manner. This difficult learning process is to be supported especially for self-studies by context-sensitive feedback as well as methods and rules for analyzing and evaluating solutions suggested by the student. In the following paper we will discuss ideas for a methods kit for interactive exercises with varying complexity and difficulty as well as for a suitable methodology for an intelligent evaluation of solutions. We will also investigate how the quality of an existing educational system on logistics can be decisively improved using AI techniques. By embedding a rule-based diagnosis module that evaluates the student’s knowledge on the basis of a viewpoint description, the student is integrated into the dialogue with the educational system in a much more active way.

Return to Contents

Application of Fuzzy Logic Techniques in the BSS1 Tutoring System

KAI WARENDORF AND SU JEN TSAO

School of Applied Science, Nanyang Technological University

Nanyang Avenue, Singapore 639798

The Brilliant Scholar Series 1 (BSS1) is a tutoring system currently used by several thousand home and school users in the learning of curricular subjects such as mathematics and sciences. It is an AI-based tutoring system using heuristics to interact with users and to monitor their progress. It is believed that the use of fuzzy logic techniques can improve the performance of such tutoring systems, by introducing intelligent features which can better manage the student’s learning. A general fuzzy logic engine was designed and implemented to support development of intelligent features for BSS1. In order to develop such features, the problem had to be suitably modeled and a knowledge base created, followed by testing and tuning with appropriate procedures.

The usefulness of such a fuzzy system depends on the engineer’s ability to model the problem suitably, define fuzzy variables and suitable membership functions for their fuzzy sets, and develop a comprehensive set of rules relating input and output variables. The average engineer who may not be equipped with this knowledge is still able to design (simple) systems by just manipulating the fuzzy set functions and the rules. Internal parameters which are generally provided by the user of the engine (the expert system designer) are fixed in this application by choosing widely used methods which have proved effective and are commonly referred to in the literature.

Return to Contents