Abstract | ||
---|---|---|
The authors report on the progress that has been made at Dragon Systems in speaker-independent large-vocabulary speech recognition using speech from DARPA's Wall Street Journal corpus. First they present an overview of the recognition and training algorithms. Then, they describe experiments involving two improvements to these algorithms, moving to higher-dimensional streams and using an IMELDA transformation. They also present some results showing the reduction in error rates.<> |
Year | DOI | Venue |
---|---|---|
1993 | 10.1109/ICASSP.1993.319391 | ICASSP |
Keywords | Field | DocType |
wall street journal,large vocabulary continuous speech,dragon systems,wall street journal data,imelda transformation,speech recognition,learning (artificial intelligence),training algorithm,vocabulary,training,error rate,higher dimensional stream,speaker-independent large-vocabulary speech recognition,error rates,higher-dimensional streams,speaker independent large vocabulary,continuous speech recognition,recent improvement,frequency,learning artificial intelligence,signal processing,indium tin oxide,context modeling,hidden markov models | Signal processing,Pattern recognition,Computer science,Speech recognition,Context model,Natural language processing,Artificial intelligence,Hidden Markov model,Vocabulary,Signal processing algorithms | Conference |
Volume | ISSN | Citations |
2 | 1520-6149 | 5 |
PageRank | References | Authors |
2.87 | 9 | 10 |
Name | Order | Citations | PageRank |
---|---|---|---|
Robert Roth | 1 | 46 | 27.12 |
James Baker | 2 | 66 | 15.44 |
Janet Baker | 3 | 29 | 13.90 |
Larry Gillick | 4 | 89 | 32.78 |
Melvyn Hunt | 5 | 5 | 2.87 |
Yoshiko Ito | 6 | 47 | 18.65 |
Stephen Lowe | 7 | 26 | 11.16 |
Jeremy Orloff | 8 | 17 | 6.99 |
Barbara Peskin | 9 | 68 | 25.14 |
Francesco Scattone | 10 | 65 | 15.63 |