Abstract | ||
---|---|---|
The outputs of multi-layer perceptron (MLP) classifiers have been successfully used in tandem systems as features for HMM-based automatic speech recognition. In a previous paper, we proposed Data-driven Clustered Hierarchical MLP (CHMLP) tandem system yielding improved performance by dividing the complicated global phone classification problem into simpler hierarchical tasks, in which specialized MLPs are trained to classify small clusters of confusing phones in a hierarchical structure. In this paper a bottom-up processing is further proposed to enhance the classification in the above CHMLP and offer even better performance. MLP rescoring for the tandem system is also investigated. The best result achieved 19.1% relative error reduction over the MFCC baseline. |
Year | DOI | Venue |
---|---|---|
2009 | 10.1109/ICASSP.2009.4960615 | ICASSP |
Keywords | Field | DocType |
data-driven clustered hierarchical mlp,mlp rescoring,lvcsr,improved performance,bottom-up processing,index terms— neural network,previous paper,simpler hierarchical task,tandem system,complicated global phone classification,hierarchical structure,hierarchical tandem system,better performance,hmm-based automatic speech recognition,relative error,pediatrics,bottom up,neural network,mel frequency cepstral coefficient,lattices,indexing terms,artificial neural networks,bottom up processing,clustering algorithms,speech recognition,multi layer perceptron,automatic speech recognition,feature extraction,neural networks,hidden markov models | Tandem,Mel-frequency cepstrum,Pattern recognition,Computer science,Feature extraction,Speech recognition,Artificial intelligence,Cluster analysis,Artificial neural network,Hidden Markov model,Perceptron,Approximation error | Conference |
ISSN | Citations | PageRank |
1520-6149 | 1 | 0.39 |
References | Authors | |
9 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Shuo-Yiin Chang | 1 | 13 | 1.77 |
Lin-shan Lee | 2 | 1525 | 182.03 |