"Build your own" spoken dialogue systems: automatically generating ISU dialogue systems from business user resources

Authors:
Oliver Lemon;Xingkun Liu;Helen Hastie
Affiliations:
University of Edinburgh, Edinburgh;University of Edinburgh, Edinburgh;University of Edinburgh, Edinburgh
Venue:
COLING '08 22nd International Conference on on Computational Linguistics: Demonstration Papers
Year:
2008

Citing 3
Cited 0

Grammatical Framework

Journal of Functional Programming
multithreaded context for robust conversational interfaces: Context-sensitive speech recognition and interpretation of corrective fragments

ACM Transactions on Computer-Human Interaction (TOCHI)
An ISU dialogue system exhibiting reinforcement learning of dialogue policies: generic slot-filling in the TALK in-car system

EACL '06 Proceedings of the Eleventh Conference of the European Chapter of the Association for Computational Linguistics: Posters & Demonstrations

Quantified Score

Hi-index	0.01

Visualization

Abstract

Building effective spoken dialogue systems (SDS) is currently a complex task requiring expert knowledge. Our tools give control of SDS application development to non-experts, who need only use a Graphical User Interface or GUI to develop state-of-the-art "Information State Update" (ISU) dialogue systems. Behind the GUI is a set of Advanced Dialogue Tools (ADT) that generate complete SDS based on Business User Resources. These resources include a database and a Process Model that captures the structure of an application, for example, banking or restaurant information. Also generated are speech recognition Language Models and grammars for robust interpretation of spontaneous speech. We will demonstrate how our simple GUI allows developers to easily and quickly create and modify SDS without the need for expensive speech application service providers. This demonstration shows the interface, the ADT components, and discusses some of the research issues involved. We also show an example application built with the tools: a tourist information system running on an ultra-mobile PC.