Within the Higgins project, a spoken dialogue system is being developed as a research test bed. Generally, spoken dialogue systems are not intended to handle every kind of dialogue and utterance. Instead, they are intended to manage a well-defined set of topics or areas, such as appointment booking or train table information. These topics or areas are often called the domain of the dialogue system.
This page describes the domain of the Higgins project and why it was chosen. It also presents a short example scenario from the domain.
The primary domain of Higgins is city navigation for pedestrians. This implies that the system initially needs to find out where the user is going, and then continuously update its knowledge of where the user is, based on the user's descriptions of the environment, and to deliver apropriate directions in easily digested portions.
Secondarily, Higgins is intended to be able to deliver simple guiding and information about buildings and objects in the vicinity of the user. Guiding, as used here, means replying to questions of the type "are there any attractions nearby?" The findings of this project should be applicable on other similar domains as well, such as car navigation and museum guides.
In this domain, the system is first presented with the task of finding out where the user is going. Then, the system must continuously update its knowledge of where the user is, based on the user's descriptions of the environment, and deliver apropriate directions in easily digested portions. An example scenario is included on the scenarios page.
The domain presents a number of immediate challenges, and the system should be able to:
- Represent complex properties of objects as well as relations between objects.
- Make morphological distinctions between singular/plural, definite/indefinite.
- Engage in a dialogue to find out the user’s destination and then iteratively update a hypothesis of the user’s position based on the user’s descriptions of the surroundings.
- Compute the user’s position using spatial and temporal reasoning.
- Engage in an extended dialogue in case the user’s descriptions are insufficient for determining the position.
- Generate route directions in appropriately sized chunks, providing the optimal path and using grounded concepts in the directions.
Copyright © 2002-2004 Jens Edlund, Gabriel Skantze and the members of the Center for Speech Technology
|
