Hidden Markov model-based supertagging in a user-initiative dialogue system

Jens Bäcker, Karin Harbusch
2002 International Workshop on Tree Adjoining Grammars and Related Formalisms  
In this paper we outline the advantages of deploying a shallow parser based on Supertagging in an automatic dialogue system in a call center that basically leaves the initiative with the user as far as (s)he wants (in the literature called userinitiative or adaptive in contrast to system-initiative dialogue systems ). The Supertagger relies on a Hidden Markov model and is trained with Gennan input texts. The entire design of a Hidden Markov-based Supertagger with trigrams builds the central
more » ... e of this paper. The evaluation of our German Supertagger lags behind the English one. Some of the reasons will be addressed later on. Nevertheless shallow parsing with the Supertags increases the accuracy compared to a basic version ofKoHDaS that only relies on recurrent plausibility networks. 1. The acronym KoHDaS stands for Koblenzer Help Desk with automatic Speech recognition. In the following the basic version is called KoHDaS-NN (NN stands for Neural Networks). Later on we investigate KoHDaS-ST where ST stands for SuperTagging.
dblp:conf/tag/BackerH02 fatcat:k7eaweqsivamzdpai26hmzun6e