ISSN:
1573-1995
Source:
Springer Online Journal Archives 1860-2000
Topics:
Electrical Engineering, Measurement and Control Technology
Notes:
Abstract Data-centric speech dialogues using large-vocabulary speech have been investigated at BT Laboratories for some time now. The research to data has been focused on applications such as directory access, customer identification, auto-attendant and transaction completion. General lessons have been learned which are applicable to many goal-seeking dialogues where the vocabulary of the task is defined by the data sought. Several trials have been run which have provided evidence of good dialogue design principles. This paper presents a framework for the dialogue design of such services, discussing topics such as mixing spelling with spoken entry and the issues of confirmation and offering of information. In addition, the difficulties of designing for confusions due to homophones and synonyms have also been addressed.
Type of Medium:
Electronic Resource
URL:
http://dx.doi.org/10.1023/A:1009691328645
Permalink