This paper attempts to rehabilitate student models within intelligent tutoring systems. Recently, some researchers have questioned both the need for detailed student models and the practical possibility of building them. We regard it as axiomatic that any intelligent tutoring system needs a student model. This paper suggests some practical guidelines and changes in philosophical approach which may help in building effective student models. 1. Introduction In a review of the 1987 "Artificial Intelligence and Education" Conference, Sandberg (1987) summarised a general opinion that "detailed user models do not necessarily enhance the capability of an intelligent tutoring system ... good teaching can do without a detailed user model, because in good teaching serious misconceptions are avoided, and errors will be repaired on the spot ... it is debatable whether the cost of constructing very 3 detailed, complex user models that are runnable and have to be maintained all the time is worth...

It is now possible to build machine learning systems whose behavior is consistent with data from human students. How can education use such simulated students? Applications that help three user groups are discussed. Teachers can practice the art of tutoring byhaving them teach a simulated student. Using a simulation instead of a real student allows teachers to see how their actions affect that student's knowledge, to undo their actions, and to try their skills on students with varying prior knowledge and learning strategies. Students can learn in collaboration with a simulated student. Because the simulated student can be simultaneously an expert and a colearner, it can scaffold and guide the human's learning in subtle ways. Instructional developers can test their instruction on simulated students. Unlike formativeevaluations with real students, a simulation-based evaluation can indicate exactly what piece of the instruction caused which pieces of knowledge, and thus help developers troubleshoot their instructional designs early in the design process. For each of these three areas of application, inherent technical limitations, existing systems and prospective systems are discussed.

A critical component of model-based intelligent tutoring systems is a mechanism for capturing the conceptual state of the student, which enables the system to tailor its feedback to suit individual strengths and weaknesses. To be useful such a modeling technique must be practical, in the sense that models are easy to construct, and effective, in the sense that using the model actually impacts student learning. This research presents a new student modeling technique which can automatically capture novel student errors using only correct domain knowledge, and can automatically compile trends across multiple student models. This approach has been implemented as a computer program, ASSERT, using a machine learning technique called theory refinement, which is a method for automatically revising a knowledge base to be consistent with a set of examples. Using a knowledge base that correctly defines a domain and examples of a student's behavior in that domain, ASSERT models student errors by c...

A modeling system may be required to predict an agent’s future actions under constraints of inadequate or contradictory relevant historical evidence. This can result in low prediction accuracy, or otherwise, low prediction rates, leaving a set of cases for which no predictions are made. A previous study that explored techniques for improving prediction rates in the context of modeling students ’ subtraction skills using Feature Based Modeling showed a tradeoff between prediction rate and predication accuracy. This paper presents research that aims to improve prediction rates without affecting prediction accuracy. The FBM-C4.5 agent modeling system was used in this research. However, the techniques explored are applicable to any Feature Based Modeling system, and the most effective technique developed is applicable to most agent modeling systems. The default FBM-C4.5 system models agents’ competencies with a set of decision trees, trained on all historical data. Each tree predicts one particular aspect of the agent’s action. Predictions from multiple trees are compared for consensus. FBM-C4.5 makes no prediction when predictions from different trees contradict one another. This strategy trades off reduced prediction rates for increased accuracy. To make predictions in the absence of consensus, three techniques have been evaluated. They include using voting, using a tree quality measure and using a leaf quality measure. An alternative technique that merges multiple

Abstract: When students attempt to solve problems collaboratively in learning environments they may miss opportunities to use available resources for achieving learning goals. We present an approach to qualitative analysis of such "missed opportunities " ("MOs") in collaborative problem-solving interactions, and discuss how the analysis can contribute to the design of the "CHENE " Computer Supported Collaborative Learning ("CSCL") system, that is used to support physics modelling tasks. Since benefits of collaboration require involvement of both partners, we concentrate on MOs to use one's partner as a resource in achieving goals of coconstructing domain concepts. After presenting analyses of different cases of MOs of this type, we discuss why MOs occur and how they may be identified. In conclusion, we propose a "minimal graded intervention " approach to guidance in CSCL environments that is intended to address the problem of MOs for learning.

: The evolution of intelligent tutoring systems (ITS) toward the use of multiple learning strategies calls on a multi-agent architecture. We show how such an architecture can be defined for the pedagogical component of an ITS. After considering the evolution of intelligent agents, we observe that ITS need to have cognitive agents able to model the human behavior in learning situations. They represent what we call actors, a category of reactive, adaptive, instructable and cognitive agents.To assume these properties we present an actor architecture with different layers of cognition: a reactive layer, a control layer, and a cognitive layer which contains learning capabilities. We provide a detailed view of this architecture and show how it functions with an example involving the different actors of a new learning strategy, the learning by disturbing strategy. Key-words: architecture, agents, actors, cognition, learning-by-disturbing, behavior, troublemaker 2 1. Introduction Learning...

This paper presents a comprehensive conceptual framework and notation for learner modelling in intelligent tutoring systems. The framework is based upon the computational distinction between behaviour, behavioural knowledge, and conceptual knowledge (in a 'vertical' dimension) and between the system, the learner, and the system's representation of the learner (in a 'horizontal' dimension). All existing techniques for learner modelling are placed within this framework. Methods for establishing the search space for learner models and for carrying out the search process are reviewed. The framework makes clear where particular learner modelling techniques are focussed and shows that they are often complementary since they address different parts of the framework. 1 A FRAMEWORK FOR LEARNER MODELLING Pierre Dillenbourg and John Self 1. THE FRAMEWORK Learner models are important within computer-based systems intended to promote learning because they provide the means to support intelligen...

The development of intelligent tutoring systems has long been the focus of applying artificial intelligence and cognitive science in education. A new breed of intelligent learning environments called learning companion systems was developed over a decade ago. In contrast to an intelligent tutoring system, in which a computer mimics an intelligent tutor, the learning companion system assumes two roles, one as an intelligent tutor and another as a learning companion. Motivated by recent interest in agent research and other technologies, this learning companion field has received increasing attention. This study addresses issues that arise from different perspectives on this research effort. With a view to future networked learning environments, the learning companion is redefined for application to a wide spectrum of educational agent research. Accordingly, several subjects that relate to educational agents, and hence learning companions, are identified.

After their beginnings in computer-aided instruction, automated tutors have re-emerged as intelligent tutoring systems. These intelligent tutors have obtained considerable success by using results from cognitive psychology and artificial intelligence to permit non-traditional instruction which is tailored to their individual students. The success of these automated tutors is due to their precise understanding and modeling of both the student and the domain being taught. A common measure of the robustness of an automated tutor is the size of the domain that it can understand. The schema-based Prolog tutor described in this dissertation is capable of recognizing a larger class of programs than existing Prolog tutors. By using powerful generalized transformations, our Prolog tutor can generate this class of programs from a very small set of normal form programs. Thus, our Prolog tutor recognizes a larger class of programs using fewer normal form programs than existing Prolog tutors. One o...

: Input-Output Agent Modelling (IOAM) is an approach to modelling an agent in terms of relationships between the inputs and outputs of the cognitive system. This approach, together with one of the leading inductive learning algorithm, C4.5, has been adopted to build a C4.5-IOAM subtraction modeller, which aims to model students' competencies on elementary subtraction skills. Results showed that C4.5-IOAM could achieved reasonably high predictive power for this purpose. Very little attempt has been made for optimising the current system that improvement of its performance could be achieved by employing strategies and techniques for this purpose. This paper reports an experiment that studied how the system's performance could be improved with techniques of confining training examples and resolving conflicting predictions. Results show that these strategies improve the system's performance in the aspects of capturing more student errors and achieving higher prediction rate. 1 Introduction...