Robert J Ross

Situated Dialogue Systems Robert J Ross

Publisher Marketing: For 50 years the natural language interface has tempted and challenged researchers and the public in equal measure. As advanced domains such as robotic systems mature over the next ten years, the need for effective language interfaces will become more significant as the disparity between physical and language ability becomes more evident. Natural language conversation with robots and other situated systems will not only require a clear understanding of theories of language use, models of spatial representation and reasoning, and theories of intentional action and agency - but will also require that all of these models be made accessible within tractable dialogue processing frameworks. While such issues pose research questions which are significant, particularly when we consider them in the light of the many other challenges in language processing and spatial theory, the benefits of competence in situated dialogue to the fields of robotics, geographic information systems, game design, and applied artificial intelligence cannot be underestimated. This book examines the burgeoning field of Situated Dialogue Systems and describes for the first time a complete computational model of situated dialogue competence for practical dialogue systems. The book can be broadly broken down into two parts. The first three chapters examine on one hand the issues which complicate the computational modelling of situated dialogue, i.e., issues of agency and spatial language competence, and on the other hand examines theories of dialogue modelling and management with respect to the needs of the situated domain. The second part of the book then details a situated dialogue processing architecture. Novel features of this architecture include the modular integration of an intentionality model alongside an exchange-structure based organization of discourse, plus the use of a functional contextualization process that operates over both implicit and explicit content in user contributions. The architecture is described at a course level, but in sufficient detail for others to use as a starting point in their own explorations of situated language intelligence.