Deep Recurrent Neural Networks with Nonlinear Masking Layers and Two-Level Estimation for Speech Separation.
Jiantao ZhangPingjian ZhangPublished in: ICANN (4) (2019)
Keyphrases
- recurrent neural networks
- nonlinear dynamic systems
- echo state networks
- neural model
- neural network
- recurrent networks
- complex valued
- hidden layer
- feed forward
- reservoir computing
- human visual system
- multiple layers
- long short term memory
- cascade correlation
- artificial neural networks
- feedforward neural networks
- automatic speech recognition
- text to speech
- nonlinear systems
- multi layer
- speech recognition
- data mining
- real time
- speech signal
- audio visual
- chaotic time series