EduAgents

The main aim in the EduAgents research project is to carry out research on learning processes and interaction between learners and software agents. The learning environment can be either centralized or distributed. In the former a human learner studies with software agents utilizing only his/her own computer while in the latter learners and software agents are working together through the Internet network. The project aims at producing a general theoretical framework which takes into account the current results of learning theory concerning the importance of social collaboration and cognitive scaffolding. Innovative software agents like peer learning companion agents and teaching agents with modern teaching strategies are seen essential in achieving these goals.

Our current prototype for equation solving is a one-machine, centralized learning environment, comprising several teaching agents, learning companion agents and cognitive tools. In the EduAgents environment the learner solves elementary equation problems with the help of several teaching and learning companion agents. The learner has also several learning tools at his/her disposal, e.g. a computerized textbook and a reflection-playback tool. Our implementation environment is MS-Windows with an object-oriented version of the LPA-WinProlog.

Our prototype incorporates four learning companion agents and four teaching agents.  Initial user test with 14 seventh-graders have been carried out and reported in conferences and journals.  Currently work is under way to improve the machine learning capabilities of the learning companion agents.

Publications

EuroAIED paper

CALLab (in Finnish)

A brochure (in Finnish)

 



22 November 1999.
 
Pentti Hietala  (ph@cs.uta.fi)