John A. Self

Student Models in Computer-Aided Instruction.

In order to individualize instruction, a computer-aided instruction (CAI) program must maintain a “student model”, i.e. a representation of the hypothesized knowledge state of the student. We propose a student model consisting of a set of programs to represent the students knowledge state. Teaching proceeds after a comparative evaluation of student and teacher programs, and learning is represented by direct modification of the student model. The advantages of an explicit procedural model are illustrated by considering a program which maintains such a model. Theoretical and practical difficulties facing such an approach are also discussed.

Applying Interactive Open Learner Models to Learning Technical Terminology

Our work explores an interactive open learner modelling (IOLM) approach where learner diagnosis is considered as an interactive process involving both a computer system and a learner that play symmetrical (to a certain extent) roles and construct together the learner model. The paper presents an application of IOLM for diagnosing and fostering a learners conceptual understanding in a terminological domain. Based on an experimental study, we discuss computational and educational benefits of IOLM in terms of improving the quality of the obtained learner model and fostering reflective thinking.

Beyond Intelligent Tutoring Systems: Situations, Interactions, Processes and Affordances

The idea of an intelligentsystem to support learning has been epitomisedby that of an intelligent tutoring system(ITS). However, ITSs are, in fact, just aparticular kind of intelligent system tosupport learning whose components reflect thevalues of the particular view that ITSsemphasise in regard to the nature of knowledge,learning and teaching, which have led to anarchitecture that focuses on representing theknowledge to be learned (domain model),inferring the learners knowledge (learnermodel), and planning instructional steps to thelearner (teaching model). On the other hand,other views of learning may lead to differentneeds in terms of knowledge representation,reasoning, and decision making capabilities inthe intelligent systems that support them.Constructivist views, for example, emphasisedifferent values and may require an entirelydifferent architecture of intelligent system tosupport its philosophy of learning. This paperpresents an architecture of an intelligentsystem to support learning that is able toaddress the issues that arise fromconstructivist theories of learning in a waythat, rather than opposing to the standard ITSarchitecture, characterises a broader view inall its components which can be appropriatelyattuned to address the issues of particularphilosophies.

The Design and Implementation of a Graphical Communication Medium for Interactive Open Learner Modelling

Our work explores an interactive open learner modelling (IOLM) approach where a learner is provided with the means to inspect and discuss the learner model. This paper presents the design and implementation of a communication medium for IOLM. We justify an approach of inspecting and discussing the learner model in a graphical manner using conceptual graphs. Based on an empirical study we draw design recommendations, taken into account in the implementation of the communication medium in STyLE-OLM - an IOLM system in a terminological domain. The potential and improvements of the medium are discussed on the basis of study with STyLE-OLM.

LeCS: A Collaborative Case Study System

Research on Intelligent Tutoring Systems has much to contribute to distance learning applications. Recently the distance learning area has experienced an increasing demand for the development of applications that make use of the new information and communication technologies (e.g., the web) as the medium for delivering courses. Intelligent distance learning can supply this demand and also provide real-time support to the distance learning process through the incorporation of ITS techniques. In this paper we describe LeCS - a collaborative case study system. LeCS supports web-based distance learning from case studies, allowing collaborative learning between a group of learners that is geographically dispersed. It provides the necessary tools to carry out the case solution development and accomplishes functions that altogether assist the learning process.

Maintaining a Jointly Constructed Student Model

Allowing the student to have some control over the diagnosis inspecting and changing the model the system has made of him is a feasible approach in student modelling which tracks the dynamics of student behaviour and provides for reflective learning. We present an approach for maintaining the student model in interactive diagnosis where a computer and a student discuss about the students knowledge. A belief modal operator is adapted to model the knowledge of the learner and to help in maintaining the interaction between the computer system and the learner. A mechanism for finding agreements and conflicts between system and learner?s views is described.

Multi-Agent Based Pedagogical Games

MCOE (Multi-agent Co-operative Environment) is a game modelled using a multi-agent system architecture composed of a society of agents which work to achieve a common goal: to assist a student to fight against pollution resulting from foreign elements (polluters) and to maintain the equilibrium of the environment [GIR98]. The system has two kinds of agents: reactive (designed and implemented using the object-oriented approach) and cognitive (designed with a mental state approach using the agent model proposed by Mora et al. in [MOR 98]). The agent model used allows us to both formally define the cognitive agent and to execute them to verify the accuracy of the model. The elements of the tutoring agent architecture are: the tutor mental states, the knowledge about what tool to use with each specific foreign elements, the knowledge about energy levels of scenario elements, the rules to help it to decide what action should be taken according to changes of the students’ mental states; and sensors for receiving data about the environment.

A Process-Sensitive Learning Environment Architecture

In this paper we describe an architecture for intelligent learning environments that provides means to analyse the process rather than the product of learning, in order to support the development of individualised learning experiences in environments oriented by a more open and constructivist philosophy of learning. The architecture is described in terms of the conceptual structure and functionality required to support the representation and analysis of processes of learner-environment interaction and the adaptation of the environment according to the state of the learning process revealed by the analysis. An application in which a simple learning environment was implemented using the architecture exemplifies the approach.

Deconstructionist Student Models in the Computer-Based Learning of Science

Student models are controversial components of computer-based learning systems. The aim of this paper is to review various issues concerned with student modelling and their place within the design process from the point of view of four themes of contemporary thinking: rational, pragmatic, critical and radical. It is seen than many of the recent trends in student modelling research can be related to postmodern ideas about the role of technology.

MArCo: Using Meta-cognitive Conflicts to Provoke Strategic Changes

The inevitability of conflicts in group interactions and their role as promoters of cognitive changes as well as of reflection and articulation has generated a significant amount of research on the subject. In this light, we have built a computational framework for the detection and mediation of meta-cognitive conflicts, aiming at fostering more productive interactions. When mediating a conflict situation, we diagnose which strategic change is happening, and suggest courses of action that can lead to more refined solutions. In order to make mediation more effective, we have also considered group and individual models, the history of the interaction and model of the task. Our approach is exemplified by the implementation of MArCo.