Handling of user interruption to achieve timing‐free utterances for spoken dialogue interface

Hideaki Kikuchi*, Ikuo Kudo, Tetsunori Kobayashi, Katsuhiki Shirai

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

This paper considers the spoken dialogue interface, which is the basic system for the multimodal interface by speech, and discusses the handling of the interruption that ensures the user the timing‐free utterances. In the interruption by the user, if the utterance by the system is designed with the sentence as the unit, as in the conventional system, there will be produced a difference between the designed content of the utterance and the actually uttered content or the content received by the user. From such a viewpoint, the following method is proposed. The unit for the utterance planning is defined as the information to be transmitted by a sentence, and which information is received is always monitored through the management of the turn‐taking between the speakers of the dialogue. As a result of experiment, the handling of smooth interruption is recognized in more than half the subjects. Using the proposed method, the difference between the utterance planned by the system and the utterance received by the user is reduced. It is verified also that the interface is made more convenient by the handling of the interruption. The time required for the user to complete the task is reduced by 7 percent, and the number of positive topic proposals is increased by 21 percent.

Original languageEnglish
Pages (from-to)92-104
Number of pages13
JournalSystems and Computers in Japan
Volume26
Issue number14
DOIs
Publication statusPublished - 1995

Keywords

  • Spoken dialogue
  • interruption
  • turn‐taking
  • utterance timing

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Information Systems
  • Hardware and Architecture
  • Computational Theory and Mathematics

Fingerprint

Dive into the research topics of 'Handling of user interruption to achieve timing‐free utterances for spoken dialogue interface'. Together they form a unique fingerprint.

Cite this