Modular neural networks exploit large acoustic context through broad-class posteriors for continuous speech recognition

ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings

Volumn 1, Issue , 2001, Pages 505-508

  Antoniou, C ^a  

^a UNIVERSITY OF ESSEX   (United Kingdom)

Author keywords

[No Author keywords available]

Indexed keywords

BROAD CLASS POSTERIORS; PHONE POSTERIOR;

ACOUSTIC SIGNAL PROCESSING; COMPUTER SIMULATION; NEURAL NETWORKS; PROBABILITY; VECTORS;

SPEECH RECOGNITION;

EID: 0034855362     PISSN: 15206149     EISSN: None     Source Type: Conference Proceeding    
DOI: None     Document Type: Conference Paper

References (11)

1. Using modular/ensemble neural networks for the acoustic modelling of speech
2. Modular neural networks exploit multiple front-ends to improve speech recognition systems
3. Acoustic modelling using modular/ensemble combinations of heterogeneous neural networks
4. Modular neural network classifiers: A comparative study
5. Speech trajectory discrimination using the minimum classification error learning
6. ACID/HNN - Clustering hierarchies of neural networks for context-dependent connectionist acoustic modelling
7. Hierarchical mixtures of experts and the EM algorithm
8. Speaker-independent phoneme recognition using hidden Markov models
9. Improved phone recognition using Bayesian triphone models
10. Multinet: A new connectionist architecture for speech recognition
11. A recurrent error propagation network speech recognition system