HLT '91 Proceedings of the workshop on Speech and Natural Language
Automatic detection and modeling of new words in a large-vocabulary continuous speech recognition system
Phonetic classification on wide-band and telephone quality speech
HLT '91 Proceedings of the workshop on Speech and Natural Language
Automatic modeling for adding new words to a large-vocabulary continuous speech recognition system
ICASSP '91 Proceedings of the Acoustics, Speech, and Signal Processing, 1991. ICASSP-91., 1991 International Conference
Speech recognition using stochastic segment neural networks
ICASSP'92 Proceedings of the 1992 IEEE international conference on Acoustics, speech and signal processing - Volume 1
| Hi-index | 0.00 |
In this paper, we extend our work on automatic detection and addition of new words. Specifically, we extend our automatic procedure for the addition of new words to a speech recognition system to include alternate pronunciations for the new words. We investigate methods for adaptation to new words after they are added to the system. For adaptation, our goal was the improvement of the accuracy of the system on the new words, using only a limited amount of speech data. All our experiments are performed within the stochastic explicit-segment speech recognition system. We evaluated our techniques on recognizing 25 isolated city names from a speech corpus, CITRON, which was collected from real users over the telephone network. On that task, we show improvement in accuracy from 34% error rate. when trained on NTIMIT alone, to 8% after adapting to 30 tokens. on average. from each new word.