Name
Title | ||
|---|---|---|
Hybrid architectures for complex phonetic features classification: a unified approach |
Abstract | ||
|---|---|---|
This paper examines how to exploit the advantages of a hybrid approach in order to overcome the drawbacks of classic automatic speech recognition (ASR) systems faced with complex phonetic features. The key idea consists of 'boosting' the capacity of a baseline ASR system to identify features as subtle as emphasis, gemination or relevant vowel lengthening. The 'booster' part is composed of a mixture of time delay neural networks (TDNNs) using an autoregressive version of the backpropagation algorithm. We choose to carry out trials on the Arabic language, which is characterized by the presence of complex features. We use three baseline systems: hidden Markov models (HMM), optimized version of learning vector quantization algorithm (O2LVQ1) and classical K-nearest neighbors' classifier (KNN). The reported results showed clearly the effectiveness of the approach since the three hybrid systems (HMM/TDNN, O2LVQ1/TDNN, KNN/TDNN) perform significantly better than their corresponding baseline systems |
Year | DOI | Venue |
|---|---|---|
2001 | 10.1109/ISSPA.2001.950249 | Signal Processing and its Applications, Sixth International, Symposium. 2001 |
Keywords | Field | DocType |
autoregressive processes,backpropagation,feature extraction,hidden Markov models,neural nets,pattern classification,speech recognition,Arabic language,K-nearest neighbor classifier,automatic speech recognition systems,autoregressive backpropagation algorithm,complex phonetic features classification,emphasis,gemination,hidden Markov models,hybrid architectures,optimized learning vector quantization algorithm,time delay neural networks,unified approach,vowel lengthening | Pattern recognition,Computer science,Learning vector quantization,Speech recognition,Time delay neural network,Artificial intelligence,Boosting (machine learning),Classifier (linguistics),Backpropagation,Artificial neural network,Hidden Markov model,Hybrid system | Conference |
Volume | ISBN | Citations |
2 | 0-7803-6703-0 | 1 |
PageRank | References | Authors |
0.35 | 3 | 2 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Sid-Ahmed Selouani | 1 | 15 | 5.71 |
| Douglas D. O'Shaughnessy | 2 | 398 | 84.79 |