ISCA Archive Interspeech 2008
ISCA Archive Interspeech 2008

Combining statistical and syntactical systems for spoken language understanding with graphical models

S. Schwarzler, J. Geiger, J. Schenk, M. Al-Hames, B. Hornler, Günther Ruske, Gerhard Rigoll

There are two basic approaches for semantic processing in spoken language understanding: a rule based approach and a statistic approach. In this paper we combine both of them in a novel way by using statistical and syntactical dynamic bayesian networks (DBNs) together with Graphical Models (GMs) for spoken language understanding (SLU). GMs merge in a complex, mathematical way probability with graph theory. This results in four different setups which raise in their complexity. Comparing our results to a baseline system we achieve a F1-measure of 93:7% in word classes and 95:7% in concepts for our best setup in the ATIS-Task. This outperforms the baseline system relatively by 3:7% in word classes and by 8:2% in concepts. The experiments were performed with the graphical model toolkit (GMTK).


doi: 10.21437/Interspeech.2008-264

Cite as: Schwarzler, S., Geiger, J., Schenk, J., Al-Hames, M., Hornler, B., Ruske, G., Rigoll, G. (2008) Combining statistical and syntactical systems for spoken language understanding with graphical models. Proc. Interspeech 2008, 1590-1593, doi: 10.21437/Interspeech.2008-264

@inproceedings{schwarzler08_interspeech,
  author={S. Schwarzler and J. Geiger and J. Schenk and M. Al-Hames and B. Hornler and Günther Ruske and Gerhard Rigoll},
  title={{Combining statistical and syntactical systems for spoken language understanding with graphical models}},
  year=2008,
  booktitle={Proc. Interspeech 2008},
  pages={1590--1593},
  doi={10.21437/Interspeech.2008-264}
}