Context Use to Improve the Speech Understanding Processing

We developed an environment for speech understanding, based on a stochastic representation of conceptual entities. The conceptual segmentation can be performed with the Viterbi or A* algorithms. The aim of this paper is to propose a model of contextual knowledge to improve the speech understanding process. We will focus on the context provided by the system prompt and the dialogue history. We argue that the context could help the prediction of the conceptual segmentation of the utterance. So results obtained on understanding error rate and running time with and without contextual knowledge are discussed.

[1]  Frank Wessel,et al.  Robust dialogue-state dependent language modeling using leaving-one-out , 1999, 1999 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings. ICASSP99 (Cat. No.99CH36258).

[2]  Günther Görz,et al.  Robust Parsing of Spoken Dialogue Using Contextual Knowledge and Recognition Probabilities , 1995, ArXiv.

[3]  Chin-Hui Lee,et al.  A speech understanding system based on statistical representation of semantics , 1992, [Proceedings] ICASSP-92: 1992 IEEE International Conference on Acoustics, Speech, and Signal Processing.

[4]  Paolo Baggia,et al.  Specialized language models using dialogue predictions , 1996, 1997 IEEE International Conference on Acoustics, Speech, and Signal Processing.

[5]  Guy Perennou,et al.  Language modelling and spoken dialogue systems - the ARISE experience , 1999, EUROSPEECH.

[6]  Andreas Kellner,et al.  Improving speech understanding by incorporating database constraints and dialogue history , 1996, Proceeding of Fourth International Conference on Spoken Language Processing. ICSLP '96.

[7]  Bernhard Rüber,et al.  Context-dependent probability adaptation in speech understanding , 1997, Comput. Speech Lang..