Following Cohen, Levesque, Perrault, and Sadek we suppose that dialogues can be analyzed in the framework of intensional logics. Starting from a multimodal logic of beliefs and intentions we investigate how beliefs of both the speaker and the hearer can be reconstructed after a speech act. The aim is to define automated deduction methods that can be integrated into cooperative human-machine and machine-machine dialogue systems.
To that end we take profit of general results from our research on belief dynamics, in particular from the topics belief base updating, and reasoning about actions, and from our research on automated theorem proving (in particular on tableau systems).
This topic is supported by a national project financed by the R&D Center of French Telecom CNET within their scientific area `Intelligent Interaction and Dialogue'. The project is called Méthodes et techniques pour la réalisation de systèmes de dialogues homme-machine coopératifs intelligents à composante orale and aims at the development of cooperative man-machine dialog systems.
The Applied Logic Group also participates in the
IRIT Working Group
on Dialogue (pages in French).
last update: may 2000
Andreas.Herzig@irit.fr / http://www.irit.fr/~Andreas.Herzig