Speech processing systems must be able to cope with recognition errors or non grammatical input from the users. We present an approach used in Verbmobil to robustly process domain relevant information using cascaded automata. One set of automata is used to extract expressions representing relevant data for the travel planning and hotel reservation domains. This information is e.g. used by the dialogue processing module. Another set of automata is used to generate natural language expressions from these representations. This robust reductionistic translation is one of four translation tracks within Verbmobil .
Cite as: Reithinger, N. (1999) Robust information extraction in a speech translation system. Proc. 6th European Conference on Speech Communication and Technology (Eurospeech 1999), 2427-2430, doi: 10.21437/Eurospeech.1999-533
@inproceedings{reithinger99_eurospeech,
author={Norbert Reithinger},
title={{Robust information extraction in a speech translation system}},
year=1999,
booktitle={Proc. 6th European Conference on Speech Communication and Technology (Eurospeech 1999)},
pages={2427--2430},
doi={10.21437/Eurospeech.1999-533}
}