An algorithm for avoiding paradoxical arguments among the multi-agent in the discourse communicator

Abstract

We developed an Intelligent Conversational Channel as a platform for knowledge sharing. Discourse Communicator is an interactive session with multiple agents that communicate the knowledge by replicating the group discussions which had transpired earlier. The knowledge, which is created by the community, is represented using story object and multimedia objects in order to represent the implicit and explicit knowledge. The story object is written using structured and unstructured language. The structured language uses taggers in order to make its semantic identifiable. The semantic carries the meaning in terms of the types of arguments. The arguments are the parameters that are passed over to the multi-agents. The agents take turn in representing the discourse. Since two arguments can have an opposite meaning, the assignment of the argument class has to make the agents be persistent in their opinion. In this paper, we show the algorithm in re-processing the arguments in a single layer and multi-layer response network. © Springer-Verlag 2004.