Digital Signal Processing Reference
In-Depth Information
5.
CONCLUSIONS AND FUTURE WORK
In this chapter, we have addressed important issues for the development of
a multi-speaker dialogue system. The interaction types between the speakers
and the system are analyzed, and, an algorithm of the multi-speaker dialogue
management is presented. Based on the proposed techniques, an MSDS
system was built to provide vehicular navigation information and assistance
in the car environment where every passenger may want to interact with the
system. The proposed MSDS system can interact with multiple speakers and
resolve conflicting opinions. Speakers are also able to acquire multi-domain
information independently or cooperatively.
Since our research is in the initial stage, only interaction (c.f. intra-action)
between speakers is studied in this manuscript. To model both the interaction
and intra-action in an MSDS is a more difficult task and requires further
studies, both in the theoretical as well as in the practical arena. Research
concerning multi-speaker spoken dialogue systems (MSDS) is in its initial
stage and we hope that our works will help to encourage further research into
the techniques of MSDS.
As future works, we plan to investigate the communication model for both
inter-action and intra-action in an MSDS. We will try to combine blind source
separation (BSS) techniques to deal with simultaneous MSDS, i.e., to allow
speakers to utter simultaneously in order to provide a more natural and
convenient MSDS system.
REFERENCES
Young, S.J., “Talking to Machines (Statistically Speaking)”, in the Proceeding of
International Conference on Spoken Language Processing, Denver, Colorado, 2002.
Bull, M. and Aylett, M., “An Analysis of The Timing of Turn-Taking in A Corpus of
Goal-Oriented Dialogue”, in Proceedings of the International Conference on Spoken
Language Processing, volume 4, pages 1175-1178, Sydney, Australia, 1998.
Poesio, M., “Cross-speaker Anaphora and Dialogue Acts”, in Proceeding of the workshop
on Mutual Knowledge, Common
[1]
[2]
[3]
Ground and Public Information ESSLLI Summer
School, 1998.
Berg, J. and Francez, N., “A Multi-Agent Extension of DRT, .Technical report of
Laboratory for Computation Linguistics”, in Proceeding of the 1st International Workshop
on Computational Semantics, pp. 81-90. University of Tilburg, 1994.
Cohen, P.R., Coulston, R. and Krout, K., “Multiparty Multimodal Interaction: A
Preliminary Analysis”, in Proceeding of International Conference on Spoken Language
Processing, 2002.
Hinkelman, E.A. and Spaceman, S.K., “Communication with Multiple Agents”, in
Proceedings of the 15th International Conference on Computational Linguistics
(COLING'94), vol. 2, pp. 1191-1197, Kyoto, Japan, 1994.
[4]
[5]
[6]
