In human dialog a wide variety of acknowledgments are used. One function of this seems to be indicating attention, interest, and involvement to the other speaker, and we believe this is an important factor in encouraging him and keeping up his interest. Thus, in this paper we focus on the problem of choosing appropriate acknowledgments at each point. Based on study of Japanese memory game dialogs, we propose an algorithm for choosing among acknowledgment responses, including `hai' (yes), `so' (right), and `un' (mm). The primary factors involved are aspects of the speaker's internal state, including confidence and liveliness, as inferred from the context and the speaker's prosody. Evaluation of naturalness and helpfulness of dialog generated by this algorithm suggests that judges prefer rule-based responses to randomly chosen responses, confirming our hypothesis that `sensitive' and subtle choice of response may improve helpfulness and naturalness of man-machine spoken language interaction.
Cite as: Tsukahara, W. (1998) An algorithm for choosing Japanese acknowledgments using prosodic cues and context. Proc. 5th International Conference on Spoken Language Processing (ICSLP 1998), paper 0955, doi: 10.21437/ICSLP.1998-139
@inproceedings{tsukahara98_icslp,
author={Wataru Tsukahara},
title={{An algorithm for choosing Japanese acknowledgments using prosodic cues and context}},
year=1998,
booktitle={Proc. 5th International Conference on Spoken Language Processing (ICSLP 1998)},
pages={paper 0955},
doi={10.21437/ICSLP.1998-139}
}