TIMIT phoneme recognition using an HMM-derived recurrent neural network.
Les T. NilesPublished in: EUROSPEECH (1991)
Keyphrases
- recurrent neural networks
- hidden markov models
- speech recognition
- viterbi algorithm
- feed forward
- complex valued
- recurrent networks
- neural network
- echo state networks
- automatic speech recognition
- long short term memory
- phoneme recognition
- feedforward neural networks
- neural model
- hidden layer
- gesture recognition
- handwritten word recognition
- artificial neural networks
- nonlinear dynamic systems
- speech synthesis
- support vector machine svm
- reservoir computing
- expert systems