IPA: improved phone modelling with recurrent neural networks
- 1 April 1994
- proceedings article
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. i, I/37-I/40
- https://doi.org/10.1109/icassp.1994.389361
Abstract
This paper describes phone modelling improvements to the hybrid connectionist-hidden Markov model speech recognition system developed at Cambridge University. These improvements are applied to phone recognition from the TIMIT task and word recognition from the Wall Street Journal (WSJ) task. A recurrent net is used to map acoustic vectors to posterior probabilities of phone classes. The maximum likelihood phone or word string is then extracted using Markov models. The paper describes three improvements: connectionist model merging; explicit presentation of acoustic context; and improved duration modelling. The first is shown to provide a significant improvement in the TIMIT phone recognition rate and all three provide an improvement in the WSJ word recognition rateKeywords
This publication has 6 references indexed in Scilit:
- An application of recurrent nets to phone probability estimationIEEE Transactions on Neural Networks, 1994
- Learning classification treesPublished by Springer Nature ,1993
- The design for the wall street journal-based CSR corpusPublished by Association for Computational Linguistics (ACL) ,1992
- Stacked generalizationNeural Networks, 1992
- Perceptual linear predictive (PLP) analysis of speechThe Journal of the Acoustical Society of America, 1990
- Segmental durations in connected-speech signals: Current resultsThe Journal of the Acoustical Society of America, 1988