The I-TUTOR project team would like to show you the creative process behind the project work. We conducted short interviews with various contributors to offer you a unique view on the I-TUTOR development. Our second interview is with Roberto Pirrone from University of Palermo, Italy.
I lead a research group, which joined the project to accomplish both research activity in Artificial Intelligence and HCI applications for VLEs, and to develop the software tool along with the University of Reading partner.
What was your main interest in the project?
We were mainly interested to intelligent support to the interaction between the user, and the system. In the I-TUTOR project there was a challenging mix of obstacles to be overcome: the interface should have had to dialogue with users in four languages, and many atomic data had to be taken into consideration to provide teachers and tutors with useful information for monitoring and/or profiling students. At the same time, students had to elicit their meta-cognitive abilities while using the system. We adopted an almost visual interface were the textual interaction is reduced to the minimum level, and information is conveyed graphically. At the same time, we used semantic computing to allow the teacher building a concept map for representing the course domain, and indexing the learning materials of the course. The teacher builds the map by providing a textual corpus where all the relevant concepts in the course domain are described and tagged with suitable keywords. The map is represented in a zooming user interface (ZUI) where the user sees clouds of keywords, and she can zoom into a single region discovering secondary keywords. At the maximum zoom level, the user can browse the course materials directly. Materials produced by the students, and social interactions (i.e. chats and forums) are projected into the map thus producing a visualization of their “activity” in the learning environment. Maps are running currently in four languages: English, Italian, Greek, and Hungarian.
What did you find suprising, more complicated or easier than expected?
As a software engineer I was challenged by the issues posed by system integration: Java based W3C standard components have been developed and coupled with Moodle that is a PHP based framework. We addressed this problems inserting our components into some Moodle plugins, and developing purposed PHP code for managing interaction and control passing. There were some security issues but we obtained a very scalable solution, which was replicated for all the components.
What new questions arise while working on the project?
New questions are related to the possibility of building a course domain representation, which should be able to elicit deep semantics and contextual relations between the relevant concepts better than the current result. A the same time, a minimum effort should be required to the teacher when she describes the domain.
What do you think about the future of the results, what would you like to see as a continuation of the project?
The whole partnership made a good job in the project, and I think there is room for deepening the research issues that are still open. I’d like to continue the project in the direction of intelligent support to the learning design. This aspect has been not developed in deep due to the project’s time constraints. On the other hand, it is very interesting because it should be faced using the findings in the field knowledge discovery, and workflow management.