![]() |
ESCA Workshop on Spoken Dialogue SystemsVigsų, Denmark |
![]() |
Conversational approaches to cooperative human-computer interaction have mostly been developed for natural language interfaces. Recently, we observe an increasing number of conversational approaches to multimodal interfaces as well (see, e.g., [Maybury, 1993, Carbonell, 1994]). They claim that linguistic contributions to the dialogue (written or spoken) as well as graphical operations are to be interpreted as communicative acts which express discourse goals. Research, however, has concentrated on the design of multimedia presentation and explanation systems and few approach the design of multimodal information retrieval interfaces on the basis of elaborate discourse models.
As information seeking and retrieval are inherently interactive processes, a flexible and cooperative user interface has to take advantage of the interactional capabilities and preferences of the user. Users' conceptions of their information problem and hence their retrieval strategies typically change as the dialogue evolves; therefore, in order to adapt its behavior to the current situation, a cooperative system should be capable of engaging in supportive (meta-)dialogues to clarify the user's goals, explain the possibilities that the system can offer, and negotiate further information seeking strategies. While advanced user interfaces rely mainly on multimedia presentations and direct manipulation techniques, supportive meta-dialogues are usually carried out linguistically. We believe that a substantial improvement in both functionality and user acceptance of such interfaces can be achieved by the integration of spoken language.
Bibliographic reference. Bateman, John / Hagen, Eli / Stein, Adelheit (1995): "Dialogue modeling for speech generation in multimodal information systems", In SDS-1995, 225-228.