Robust speech understanding

Astrid Brietzmann, Fritz Class, Ute Ehrlich, Paul Heisterkamp, Alfred Kaltenmeier, Klaus Mecklenburg, Peter Regel-Brietzmann

This paper describes a fully operational system for analyzing all aspects of continuous speech from word recognition up to linguistic representation. Many systems rely on fully grammatical speech input, others use only shallow analysis without using declarative linguistic knowledge. We propose a flexible processing, where the depth of the analysis varies according to internal (quality of the speech input) and/or external (limitations, e.g. of processing time) criteria. In architecture and system design, special effort was made to cover effects of spontaneous speech (e.g. unknown words, pauses, or sentence breaks). The described system is part of a speech dialog system not discussed in this paper.

doi: 10.21437/ICSLP.1994-256

Cite as: Brietzmann, A., Class, F., Ehrlich, U., Heisterkamp, P., Kaltenmeier, A., Mecklenburg, K., Regel-Brietzmann, P. (1994) Robust speech understanding. Proc. 3rd International Conference on Spoken Language Processing (ICSLP 1994), 967-970, doi: 10.21437/ICSLP.1994-256

@inproceedings{brietzmann94_icslp,
  author={Astrid Brietzmann and Fritz Class and Ute Ehrlich and Paul Heisterkamp and Alfred Kaltenmeier and Klaus Mecklenburg and Peter Regel-Brietzmann},
  title={{Robust speech understanding}},
  year=1994,
  booktitle={Proc. 3rd International Conference on Spoken Language Processing (ICSLP 1994)},
  pages={967--970},
  doi={10.21437/ICSLP.1994-256}
}