Spoken language variation over time and state in a natural spoken dialog system

We are interested in adaptive spoken dialog systems for automated services. Peoples' spoken language usage varies over time for a fixed task, and furthermore varies depending on the state of the dialog. We characterize and quantify this variation based on a database of 20 K user-transactions with AT&T's experimental 'How May I Help You?' spoken dialog system. We then report on a language adaptation algorithm which was used to train state-dependent ASR language models, experimentally evaluating their improved performance with respect to word accuracy and perplexity.

[1]  Allen L. Gorin,et al.  Generating semantically consistent inputs to a dialog manager , 1997, EUROSPEECH.

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

[3]  Giuseppe Riccardi,et al.  Grammar Fragment acquisition using syntactic and semantic clustering , 1998, Speech Commun..

[4]  A. Gorin.  On automated language acquisition , 1989 .

[5]  Andrej Ljolje,et al.  A spoken language system for automated call routing , 1997, 1997 IEEE International Conference on Acoustics, Speech, and Signal Processing.

[6]  Roberto Pieraccini,et al.  Stochastic automata for language modeling , 1996, Comput. Speech Lang..

[7]  Hiroyuki Sakamoto,et al.  Continuous speech recognition using a dialog-conditioned stochastic language model , 1994, ICSLP.

[8]  Giuseppe Riccardi,et al.  How may I help you? , 1997, Speech Commun..

[9]  Giuseppe Riccardi,et al.  Automatic acquisition of salient grammar fragments for call-type classification , 1997, EUROSPEECH.