An architecture for a generic dialogue shell

Abstract

This paper describes our work on dialogue systems that can mimic human conversation, with the goal of providing intuitive access to a wide range of applications by expanding the user's options in the interaction. We concentrate on practical dialogue: dialogues in which the participants need to accomplish some objective or perform some task. Two hypotheses regarding practical dialogue motivate our research. First, that the conversational competence required for practical dialogues, while still complex, is significantly simpler to achieve than general human conversational competence. And second, that within the genre of practical dialogue, the bulk of the complexity in the language interpretation and dialogue management is independent of the task being performed. If these hypotheses are true, then it should be possible to build a generic dialogue shell for practical dialogue, by which we mean the full range of components required in a dialogue system, including speech recognition, language processing, dialogue management and response planning, built in such a way as to be readily adapted to new applications by specifying the domain and task models. This paper documents our progress and what we have learned so far based on building and adapting systems in a series of different problem solving domains.

Speech recognition in a dialog system for patient health monitoring

“…CARDIAC interprets the patient information and uses it to update its user models. The following dialogue excerpt illustrates the sort of interaction that the system supports: The CARDIAC system is built as an instantiation of the TRIPS generic dialogue architecture [6], which provides broad coverage domain-general parsing, generation, and discourse interpretation facilities. To adapt this generic system to a specific domain, we need to define a task model that the conversational agent uses to drive the interview process without needing to manage the details of natural conversation.…”

Ambient Intelligence Interaction via Dialogue Systems

“…The DM component should not be involved in the process of accessing a background system or performing domain reasoning. This divide to conquer approach assumes that practical dialogue and domain-independent hypothesis are true (Allen et al, 2000). The reason is that typical applications of human computer interaction involve dialogue focussed on accomplishing some specific task.…”

Promoting extension and reuse in a spoken dialog manager

“…The frame-based method has been used widely in a number of systems and is also used to implement mixed-initiative dialogs in VoiceXML. Finally, agent-based approaches incorporate a wide variety of approaches that use techniques from Artificial Intelligence to produce more intelligent systems, for example, by using plan-based dialog management [Allen et al 2000] or some agent-based approach [Turunen and Hakulinen 2003;Bohus and Ridnicky 2003]. The Queen's Communicator system is an example of an agent-based system as it incorporates mixed initiative interaction in which the user can change or correct items as well as switch domains, and it makes complex decisions on how to ground the dialog based on the current context and the dialog history ].…”