Sixth European Conference on Speech Communication and Technology

Budapest, Hungary
September 5-9, 1999

Automatic Speech Recognition Using Acoustic Confidence Conditioned Language Models

Richard C. Rose, Giuseppe Riccardi

AT&T Labs-Research, Shannon Laboratory, Florham Park, NJ, USA

A modified decoding algorithm for automatic speech recognition (ASR) will be described which facilitates a closer coupling between the acoustic and language modeling components of a speech recognition system. This closer coupling is obtained by extracting word level measures of acoustic confidence during decoding, and making coded representations of these confidence measures available to the ASR network during decoding. A simulation of this decoding strategy is implemented using a word lattice rescoring paradigm. A joint acoustic{language model will be described where linguistic context is augmented to include the encoded values of acoustic confidence. Finally, the performance of the word lattice based implementation of the decoding algorithm will be evaluated on a large vocabulary natural language understanding task.

Full Paper (PDF)   Gnu-Zipped Postscript

Bibliographic reference.  Rose, Richard C. / Riccardi, Giuseppe (1999): "Automatic speech recognition using acoustic confidence conditioned language models", In EUROSPEECH'99, 303-306.