A multi-domain dialog system to integrate heterogeneous spoken dialog systems

Abstract

[EN] In this paper, we present an architecture to create a multidomain
\nspoken dialog system with minimum effort by composing
\nheterogeneous pre-existent spoken dialog systems into
\na new system able to perform richer interactions. A Task Manager
\nacts as a proxy for the different sub-domains and activates
\none of the systems each turn. The different sub-systems are not
\naware that they are used in a multi-domain scenario and believe
\nthat they are speaking directly to the user. This allows us to add
\nnew domains leaving most of the underlying models unmodified
\nand reducing the amount of time and money needed to deploy
\nthe application.
\nThe proposed architecture has been applied to create a
\nmulti-domain system combining three heterogeneous spoken
\ndialog systems (sport facilities booking, weather service and
\npersonal calendar) in Spanish. The evaluation with naive real
\nusers shows that this is an appropriate approach to develop
\nmulti-domain spoken dialog systems.