AAAI Publications, 2009 AAAI Fall Symposium Series

Font Size: 
Managing Conversation Uncertainty in TutorJ
Vincenzo Cannella, Roberto Pirrone

Last modified: 2009-10-31


Uncertainty in natural language dialogue is often treated through stochastic models. Some of the authors already presented TutorJ that is an Intelligent Tutoring System, whose interaction with the user is very intensive, and makes use of both dialogic and graphical modality. When managing the interaction, the system needs to cope with uncertainty due to the understanding of the user's needs and wishes. In this paper we present the extended version of TutorJ, focusing on the new features added to its chatbot module. These features allow to merge deterministic and probabilistic reasoning in dialogue management, and in writing the rules of the system's procedural memory.


human-computer interaction; natural language interaction; uncertainty management; TutorJ

Full Text: PDF