International Journal of Artificial Intelligence in Education: The European Journal on Artificial Intelligence

The paper presents SQLT-Web, a Web-enabled intelligent tutoring system for the SQL database language. SQLT-Web is a Web-enabled version of an earlier, standalone ITS. In this paper we describe how the components of the standalone system were reused to develop the Web-enabled system. The system observes students' actions and adapts to their knowledge and learning abilities. We describe the system's architecture in comparison to the architectures of other existing Web-enabled tutors. All tutoring functions are performed on the server side, and we explain how SQLT-Web deals with multiple students. The system has been open to outside users since March 2000. SQLT-Web has been evaluated in the context of genuine teaching activities. We present the results of three evaluation studies with the University of Canterbury students taking database courses, which show that SQLT-Web is an effective system. The students have found the system a valuable asset to their learning.

MetaLinks, an authoring tool and web server for adaptive hyperbooks, is described. MetaLinks hyperbooks provide textbook-like content in a flexible manner that supports opportunistic processes of active reading and learning. Through adaptive methods a single on-line artifact can address the diverse needs of a variety of user goals and backgrounds. The system is designed to support inquiry, exploratory, or curiosity-driven learning in richly interconnected material and to ameliorate a number of usability issues that arise in such open ended learning environments. Chief among the usability issues addressed are the potential for discontinuous conceptual and narrative flow as users navigate flexibly and idiosyncratically through hyperbook content. Our goals are addressed through a number of interface features and adaptive methods including narrative smoothing, custom depth control, and thematic links. In this paper we describe the MetaLinks system, discuss a number of hyperbook design issues, and give the results of two formative evaluations of a MetaLinks Hyperbook.

One problem that commonly faces hypertext users, particularly in educational situations, is the difficulty of identifying pages of information most relevant to their current goals or interests. In this paper, we discuss the technical feasibility and the utility of applying machine learning algorithms to generate personalised adaptation on the basis of a user's browsing history in hypertext, without additional input from the user. In order to investigate the viability of this approach, we developed a Web-based information system called MLTutor. The design of MLTutor aims to remove the need for pre-defined user profiles and replace them with a dynamic user profile-building scheme in order to provide individual adaptation. In MLTutor, this adaptation is achieved by a combination of conceptual clustering and inductive machine learning algorithms. An evaluation technique that probes the detailed effectiveness of the adaptation is presented. The use of dynamic user profiles has been shown to be technically feasible; however, while a superficial evaluation indicates that it is educationally effective, the more thorough evaluation performed here shows that the positive results may be attributed to other causes. This demonstrates the need for thorough evaluation of adaptive hypertext systems.

This paper describes the design and evaluation of COLER, a coach that helps students collaborate while solving Entity Relationship modelling problems. Unlike previous work generally emphasizing dialogue analysis or expert models, this work evaluates a new approach to supporting collaboration that identifies learning opportunities based on differences between problem solutions and tracking levels of participation. The contribution made by these and other knowledge sources in the generation of collaboration advice was evaluated by comparing expert rankings of advice to the software coach's rankings, and by identifying the advice that would be lost if each respective knowledge source were removed. Results show that good quality advice can be obtained through these knowledge sources, although other knowledge sources may fill in gaps relative to the expert's performance. This work demonstrates how intelligent agents can produce reasonable collaboration advice in domains for which structured problem solutions exist by using a few basic knowledge sources, and illustrates several methods of evaluating the knowledge and reasoning of complex knowledge-based system.