Experiences from the development of Augusta multi-modal spoken dialogue system

This paper describes experiences from the development of a Swedish spoken dialogue system with a talking agent, August. The system was exposed to the general public at a very early stage. Apart from describing the system and its components this paper discusses the problems encountered during the set-up and describe possible solutions considered. The system has been used for a period of six months to collect spontaneous speech data, largely from people with no previous experience of speech technology. One of the goals of the August project was to be able to analyze how novice users interact with a multi-modal information kiosk, placed without supervision in a public location. Another goal was to demonstrate how the speech technology modules developed at the department could be put together to rapidly prototype a multi-modal spoken dialogue system.