Abstract | ||
---|---|---|
In this paper, we propose a novel method for rapid feature space Maximum Likelihood Linear Regression (FMLLR) speaker adaptation based on bilinear models. When the amount of adaptation data is limited, the conventional FMLLR transforms can be easily over-trained and can even degrade the performance. In such cases, usually by introducing structural constraints on the FMLLR transformation, the original FMLLR adaptation method can be modified for rapid adaptation. The objective of our bilinear model is to introduce a prior knowledge analysis on the training speakers based on Singular Vector Decomposition (SVD), and to incorporate it in the decoding process. This can effectively reduce the number of free parameters of FMLLR transformation and achieve performance improvements even with limited adaptation data. The efficiency of the proposed algorithm is demonstrated with experiments on the Mandarin digital dataset and the Mandarin voice search dataset respectively. |
Year | DOI | Venue |
---|---|---|
2011 | 10.1109/ICASSP.2011.5947342 | ICASSP |
Keywords | Field | DocType |
rapid speaker adaptation,svd,singular vector decomposition,decoding process,fmllr,maximum likelihood estimation,fmllr adaptation method,rapid feature space mllr speaker adaptation,speaker recognition,mandarin voice search dataset,rapid feature space maximum likelihood linear regression speaker adaptation,transforms,mandarin digital dataset,singular value decomposition,bilinear models,hidden markov model,matrix decomposition,mathematical model,data models,computer model,data model,computational modeling,hidden markov models,indexing terms,feature space | Singular value decomposition,Feature vector,Pattern recognition,Computer science,Vector decomposition,Matrix decomposition,Speech recognition,FMLLR,Speaker recognition,Artificial intelligence,Hidden Markov model,Bilinear interpolation | Conference |
Volume | Issue | ISSN |
null | null | 1520-6149 E-ISBN : 978-1-4577-0537-3 |
ISBN | Citations | PageRank |
978-1-4577-0537-3 | 0 | 0.34 |
References | Authors | |
7 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Shilei Zhang | 1 | 57 | 9.81 |
Peder A. Olsen | 2 | 398 | 37.80 |
Yong Qin | 3 | 161 | 42.54 |