AMANDA- An Intelligent Dialog Coordination Environment

This paper describes AMANDA2- an intelligent system intended to coordinate collective dialog sessions in distance learning environments. The overall objective of AMANDA is to help tutors achieve better results from group discussions and improve knowledge transfer among the participants. This is done by integrating the collective dialog as a disciplined and well-coordinated activity in distance learning situations. For this purpose, the dialog is represented as an argumentation tree, a structured collection of questions, alternatives and arguments which evolves along sequential dialog cycles. The intelligent behavior of the system is due to its coordination actions taken in response to reasoning over the dialog. We describe how AMANDA coordinates the dialog process by generating a sequence of dialog cycles based on a set of coordination parameters. In this paper we briefly describe AMANDA’s functional modules, internal structures and coordination algorithms. The knowledge models that support system reasoning are described, as well as our practical experience in domain modeling. We have tested the system in actual training situations, for which we chose a test course and modeled the corresponding domain knowledge. Although some modules of the system are still under development, specially those related to semantic reasoning, we discuss the application of semantic parameters and identify some techniques which may improve the coordination algorithm