Speaker-aware training of LSTM-RNNS for acoustic modelling.
Tian TanYanmin QianDong YuSouvik KunduLiang LuKhe Chai SimXiong XiaoYu ZhangPublished in: ICASSP (2016)
Keyphrases
- recurrent neural networks
- feedforward neural networks
- recurrent networks
- long short term memory
- echo state networks
- acoustic models
- speech recognition
- feed forward
- speaker independent
- training phase
- supervised learning
- training set
- pattern recognition
- prosodic features
- acoustic features
- speaker identification
- training process
- automatic speech recognition
- training examples
- training samples
- multi modal
- genetic algorithm
- audio visual
- data sets
- test set
- speaker verification
- hidden markov models
- active learning
- artificial neural networks
- acoustic signal