Figure-Ground Relations in Human-Computer Dialogue Modeling

Abstract

This article is devoted to cognitive aspects of establishing figure-ground relations in view of developing virtual interlocutors capable of maintaining a dialogue with humans. The nature of figure-ground recognition is reinterpreted from the perspective of computer as an active participant of dialogue interaction with the aim of improvement of its communicative competence. The results of the research suggest the ways of optimizing the functioning of automatic dialogue systems. The components which are crucial to the establishment of figure-ground relations have been modeled for virtual interlocutors, implementation of which in the development of the linguistically competent chat-bot made it possible to achieve the natural and coherent dialogical communication between the human and the machine.