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

Author

Gorin, Allen L. ; Riccardi, Giuseppe

Author_Institution

Dept. of Speech Res., AT&T Bell Labs., Florham Park, NJ, USA

Volume

2

fYear

1999

fDate

15-19 Mar 1999

Firstpage

721

Abstract

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

Keywords

interactive systems; natural languages; speech recognition; AT&T; automated services; automatic speech recognition; database; interactive speech systems; language adaptation algorithm; natural spoken dialog system; performance; perplexity; spoken language variation; state-dependent ASR language models; user-transactions; word accuracy; Adaptive systems; Automatic speech recognition; Humans; Laboratories; Natural languages; Navigation; Robustness; Speech recognition; Telephony; Transaction databases;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1999. Proceedings., 1999 IEEE International Conference on

Conference_Location

Phoenix, AZ

ISSN

1520-6149

Print_ISBN

0-7803-5041-3

Type

conf

DOI

10.1109/ICASSP.1999.759768

Filename

759768