Abstract | ||
---|---|---|
This paper describes the system and algorithmic developments in the automatic transcription of Mandarin broadcast speech made at IBM in the second year of the DARPA GALE program. Technical advances over our previous system include improved acoustic models using embedded tone modeling, and a new topic-adaptive language model (LM) rescoring technique based on dynamically generated LMs. We present results on three community-defined test sets designed to cover both the broadcast news and the broadcast conversation domain. It is shown that our new baseline system attains a 15.4% relative reduction in character error rate compared with our previous GALE evaluation system. And a further 13.6% improvement over the baseline is achieved with the two described techniques. |
Year | DOI | Venue |
---|---|---|
2008 | 10.1109/ICASSP.2008.4518613 | Las Vegas, NV |
Keywords | Field | DocType |
natural language processing,speech recognition,DARPA GALE program,IBM Gale Mandarin transcription system,Mandarin broadcast speech,acoustic models,algorithmic developments,automatic transcription,broadcast conversation domain,character error rate,embedded tone modeling,relative reduction,topic-adaptive language model rescoring technique,speech processing,speech recognition,tone modeling,topic adaptation | Speech processing,Broadcasting,IBM,Evaluation system,Computer science,Word error rate,Speech recognition,Natural language processing,Artificial intelligence,Baseline system,Mandarin Chinese,Language model | Conference |
ISSN | ISBN | Citations |
1520-6149 E-ISBN : 978-1-4244-1484-0 | 978-1-4244-1484-0 | 8 |
PageRank | References | Authors |
0.72 | 6 | 8 |
Name | Order | Citations | PageRank |
---|---|---|---|
Selina M. Chu | 1 | 8 | 0.72 |
Hong-Kwang Kuo | 2 | 71 | 9.60 |
Lidia Mangu | 3 | 1203 | 125.73 |
Yi Liu | 4 | 116 | 22.97 |
Yong Qin | 5 | 161 | 42.54 |
Qin Shi | 6 | 61 | 10.77 |
Shilei Zhang | 7 | 57 | 9.81 |
Hagai Aronowitz | 8 | 240 | 22.95 |