The Waxholm dialog project
Waxholm was a spoken dialogue system for retrieving information about the ferryboat services in the Stockholm archipelago.
Waxholm was a generic system in which speech synthesis and speech recognition can be studied in a man-machine dialogue framework. In addition, the system should facilitate the collection of speech and text data that are required for the development of the system. The demonstrator application, that we call WAXHOLM, gives information on boat traffic in the Stockholm archipelago. A fleet of some twenty boats from the Waxholm company connect about two hundred ports. Different days of the week have different time tables.
The system also had some information about facilities like hotels and restaurants on the islands. The system could be used to find information, not to carry out the actual bookings or provide further tourist information. The system featured a graphical interface with an animated talking head and a picture that visualized the system’s domain (Bertenstam et al. 1995). Textual information was presented by placing tables by the icons depicting the corresponding facilities. Information provided by the user was also displayed at different places, the recognized destination was shown on the island and the recognized departure on the jetty.
The Waxholm project was initiated in 1992 as a research effort for building spoken dialogue systems. In this project, new dialogue management and parsing modules were developed and combined with TMH’s existing speech synthesis and recognition. The goal was to acquire knowledge on how to develop the natural language modules and the other system modules needed to build spoken dialogue systems (Carlson 1996). Another important purpose was to collect spoken dialogue data. The fully automated Waxholm system has not been used in any extensive users studies. These were instead carried out during the iterative development of the system. The WOZ version was used to collect dialogues from 66 subjects. Before interacting with the system, the subjects of the WOZ experiment were given written scenarios, but these turned out to influence the subjects. Some of the subjects reused large parts of the written scenarios when they approached the system.
Group: Speech Communication and Technology
Antonio de Serpa-Latio
Duration: 1991 - 1995
Developing multimodal spoken dialogue systems. Empirical studies of spoken human-computer interaction. Doctoral dissertation, KTH. [pdf] (2002).
Proc of Twente Workshop on Language Technology. Dialogue Management in Natural Language Systems (TWLT 11) (pp. 209-218). [pdf] (1996). The dialog component in the Waxholm system. In LuperFoy, S., Nijholt, A., & Veldhuijzen van Zanten, G. (Eds.),
Phonetica Pragensia IX. Charisteria viro doctissimo Premysl Janota oblata. Acta Universitatis Carolinae Philologica 1, 1996. [pdf] (1996). The Waxholm spoken dialogue system.
Proc of ICSLP-96, 4th Intl Conference on Spoken Language Processing (pp. 677-680). Philadelphia, USA. [pdf] (1996). Generic and domain-specific aspects of the Waxholm NLP and Dialog modules. In
Proc of ESCA Workshop on Spoken Dialogue Systems (pp. 281-284). Vigs¿, Denmark. [pdf] (1995). The Waxholm system - a progress report. In Dalsgaard, P. (Ed.),
Proceednings Eurospeech 1995 (pp. 833-836). Madrid. [pdf] (1995). The Waxholm application data-base. In Pardo, J. (Ed.),
STL-QPSR, 36(1), 049-074. [pdf] (1995). Spoken dialogue data collected in the Waxholm project.
STL-QPSR, 36(1), 029-048. [pdf] (1995). The natural language component - STINA.
Proc of ESCA Workshop on Spoken Dialogue Systems (pp. 137-140). Vigs¿, Denmark. [pdf] (1995). Dialogue management in the Waxholm system. In Dalsgaard, P. (Ed.),
FONETIK -94, Working papers 43, Papers from the Eighth Swedish Phonetics Conference (pp. 22-23). Dept. of Linguistics, Lund University, Lund. [pdf] (1994). Speech recognition in the Waxholm dialog system. In
FONETIK «94, Papers from the 8th Swedish Phonetics Conference (pp. 46-49). Lund, Sweden. (1994). Dialog management in the Waxholm system. In
Proceedings of Seventh Swedish Phonetics Conference, RUUL 23 (pp. 49-52). Uppsala. [pdf] (1993). An experimental dialog system: WAXHOLM. In
Proceedings Eurospeech '93 (pp. 1867-1870). Berlin. (1993). An experimental dialogue system: WAXHOLM. In
STL-QPSR, 34(2-3), 015-020. [pdf] (1993). An experimental dialogue system: Waxholm.
Fonetik '92, the Sixth Swedish Phonetics Conference held in Gothenburg, Technical Report no. 12 (pp. 23-26). Chalmers University of Technology, Göteborg. [pdf]
(1992). STINA: A probabilistic parser for speech recognition. In Huber, D. (Ed.),