Dialog management in the Waxholm system
Rolf Carlson, Sheri Hunnicutt · 2009
In this paper we will describe the natural language and dialog component in the Waxholm system. Our parser, STINA, is knowledge based and is designed as a probabilistic language model. The dialog management, also implemented in STINA, is based on grammar rules and lexical semantic features. The parser is running with two different time scales corresponding to the words in each utterance and to the turns in the dialog. Topic selection is accomplished based on probabilities calculated from user initiatives. 1. INTRODUCTION Our research group at KTH is currently building a generic system in which speech synthesis and speech recognition can be studied in a man-machine dialog framework. The demonstrator application, which we call WAXHOLM, gives information on boat traffic in the Stockholm archipelago [1, 2]. In addition to boat time-tables, the database also contains information about port locations, hotels, camping places, and restaurants in the Stockholm archipelago. The system is prese...