Local Login

Current Projects

LORNET (press release)

The Learning Object Repositories Network (LORNET) project is a Natural Science and Engineering Research Council (NSERC) Research Network intent on identifying, formalizing, organizing, and sustaining learning technologies to achieve e-learning. It is made up of several different academic institutions and a number of industry partners organized around six different themes. The University of Saskatchewan, through the Laboratory for Advanced Research in Intelligent Educational Systems (ARIES), is leading the third theme entitled Active and Adaptive Learning Objects.

Past Projects


Teachers are not always able to provide all students with the help that they need. One way of decreasing the load on teachers is for students to help each other (peer help). Similarly, in workplace environments, there is a huge need for workers to keep up to date in a constantly changing corporate environment, but seldom enough resources for large scale training and virtually nothing at all to support an immediate help need.  It would be ideal if the workers could help each other. Unfortunately, in a large class or a big organization, students or workers may not know who to ask. Under the auspices of the TeleLearning NCE, the I-Help project was initiated in order to address this important issue. The "I" (in I-Help) stands for many things, such as "intelligent", "interactive", and "integrated", but most importantly it is the first person singular. People who receive help also give help.

The two main components in I-Help are derived from earlier ARIES research projects.  The first of these is the public discussion forum, where learners contribute to a subject-oriented discussion forum and a moderated FAQ list to provide electronic help. The public discussion forum can be accessed easily from within I-Help.  The public discussion forum derives from the CPR project, discussed separately in more detail below.  The other main component of I-Help is the one to one system, derived from the PHelpS system, developed some years ago (as the first TeleLearning project) in the workplace environment of the Prairie Regional Pyschiatric Centre of the Canadian prison system to provide help to workers learning to use a new database system.

PHelpS was able to provide an appropriately "ready, able and willing" peer helper to provide human help to anybody needing help in accomplishing tasks using this database. The selection of the appropriate help resource is based on modeling learner knowledge and on a conceptual model of the subject material.  The system accepts and interprets the help request from the learner and locates the appropriate help resource.   I-Help is breaking new ground over its predecessor systems in several ways.  Most interestingly, perhaps, is that it is built on a multi-agent architecture, also designed in the ARIES Laboratory. In this architecture, each user is represented by a "personal agent" who can negotiate both the giving and getting of help for the user. Agent-agent interactions in this architecture are governed by an economic system based on ICU's (I-Help Credit Units) exchanged between helpee and helper.

Interesting research issues arise in areas like relativistic and active user modelling, open learner modelling, distributed systems, and new pedagogies for just-in-time learning. Spin-off research projects carried out by graduate students (see below) are providing tools to help the helper give more timely and effective help, trying various agent negotiation strategies, and supporting role-based collaboration.

Currently two I-Help system has been deployed at the University of Saskatchewan for in all undergraduate computer science courses, as well as a number of non computer science classes.  This deployment is considered wide scaled, and brings support to more than 2,000 students.


The AROMA Shell for building granularity-based advisors was constructed primarily under the auspices of the IRIS Network of Centres of Excellence. AROMA is a knowledge representation and reasoning system that enables the user to specify knowledge related by the granularity relations of aggregation and abstraction. The granularity formalism provides a natural opportunistic recognition algorithm that will recognize aggregate and abstract entities from lower-level observations. Central to the granularity-based reasoning approach is the use of context and local control information to guide and restrict search, and the use of clusters of components and constraints to facilitate propagation of knowledge within the granularity hierarchies. These 4C's, context, controls, clusters, and constraints make the granularity-based recognition approach very robust and powerful. In conjunction with this recognition engine, a case-based retrieval system has been incorporated into AROMA to enable the system to efficiently match library cases to newly recognized patterns. Finally, AROMA provides a set of tools for building advising systems using this granularity engine. A hierarchy engineering tool (HE), a task engineering tool (TE), and a student interface have been developed. In addition, debugging tools for the HE and TE have been developed.