Abstract | ||
---|---|---|
The Gaussian mixture models (GMM) has proved to be an effective probabilistic model for speaker verification, and has been widely used in most of state-of-the-art systems. In this paper, we introduce a new method for the task: that using Ad- aBoost learning based on the GMM. The motivation is the fol- lowing: While a GMM linearly combines a number of Gaus- sian models according to a set of mixing weights, we believe that there exists a better means of combining individual Gaus- sian mixture models. The proposed AdaBoost-GMM method is non-parametric in which a selected set of weak classifiers, each constructed based on a single Gaussian model, is optimally combined to form a strong classifier , the optimality being in the sense of maximum margin. Experiments show that the boosted GMM classifier yields 10.81% relative reduction in equal error rate for the same handsets and 11.24% for different handsets, a significant improvement over the baseline adapted GMM sys- tem. |
Year | Venue | Keywords |
---|---|---|
2003 | INTERSPEECH | gaussian mixture model,probabilistic model |
Field | DocType | Citations |
AdaBoost,Pattern recognition,Existential quantification,Computer science,Word error rate,Speech recognition,Gaussian,Artificial intelligence,Gaussian network model,Statistical model,Classifier (linguistics),Mixture model | Conference | 3 |
PageRank | References | Authors |
0.43 | 6 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Stan Z. Li | 1 | 8951 | 535.26 |
Dong Zhang | 2 | 125 | 17.08 |
Chengyuan Ma | 3 | 119 | 12.00 |
Heung-Yeung Shum | 4 | 10824 | 739.88 |
Eric Chang | 5 | 625 | 49.79 |