Data-driven lexicon expansion for Mandarin broadcast news and conversation speech recognition Export

@inproceedings{citeulike:5907651, abstract = {We present a data-driven framework for expanding the lexicon to improve Mandarin broadcast news and conversation speech recognition. The lexicon expansion includes the generation of pronunciation variants for frequent words and vocabulary augmentation with new words and phrases derived from the training data. To learn multiple pronunciations, we first generate all possible pronunciation candidates for a word from its character pronunciation network. The top pronunciation variants are then selected from forced alignment statistics. To augment the acoustic vocabulary, we propose an efficient algorithm that derives new words based on N-gram statistics. Experiments show that a dictionary expanded in this manner yields significant improvements on a Mandarin broadcast speech recognition task.}, address = {Washington, DC, USA}, author = {Lei, Xin and Wang, Wen and Stolcke, Andreas}, booktitle = {ICASSP '09: Proceedings of the 2009 IEEE International Conference on Acoustics, Speech and Signal Processing}, citeulike-article-id = {5907651}, citeulike-linkout-0 = {http://portal.acm.org/citation.cfm?id=1583139}, citeulike-linkout-1 = {http://dx.doi.org/10.1109/ICASSP.2009.4960587}, citeulike-linkout-2 = {http://www.speech.sri.com/people/leixin/papers/LearnDict.pdf}, doi = {10.1109/ICASSP.2009.4960587}, isbn = {978-1-4244-2353-8}, keywords = {dictionary}, pages = {4329--4332}, posted-at = {2009-10-07 22:18:48}, priority = {0}, publisher = {IEEE Computer Society}, title = {Data-driven lexicon expansion for Mandarin broadcast news and conversation speech recognition}, url = {http://dx.doi.org/10.1109/ICASSP.2009.4960587}, year = {2009} }

TY - CONF ID - citeulike:5907651 L3 - citeulike-article-id:5907651 N2 - We present a data-driven framework for expanding the lexicon to improve Mandarin broadcast news and conversation speech recognition. The lexicon expansion includes the generation of pronunciation variants for frequent words and vocabulary augmentation with new words and phrases derived from the training data. To learn multiple pronunciations, we first generate all possible pronunciation candidates for a word from its character pronunciation network. The top pronunciation variants are then selected from forced alignment statistics. To augment the acoustic vocabulary, we propose an efficient algorithm that derives new words based on N-gram statistics. Experiments show that a dictionary expanded in this manner yields significant improvements on a Mandarin broadcast speech recognition task. SN - 978-1-4244-2353-8 AD - Washington, DC, USA EP - 4332 TI - Data-driven lexicon expansion for Mandarin broadcast news and conversation speech recognition PB - IEEE Computer Society T2 - ICASSP '09: Proceedings of the 2009 IEEE International Conference on Acoustics, Speech and Signal Processing SP - 4329 KW - dictionary AU - Lei, Xin AU - Wang, Wen AU - Stolcke, Andreas PY - 2009/// UR - http://dx.doi.org/10.1109/ICASSP.2009.4960587 ER -