Semi-supervised training for bottle-neck feature based DNN-HMM hybrid systems.
Haihua XuHang SuChng Eng SiongHaizhou LiPublished in: INTERSPEECH (2014)
Keyphrases
- hybrid systems
- semi supervised
- bottle neck
- training process
- hidden markov models
- neural network
- supervised learning
- black box
- intelligent systems
- discrete event
- meta learning
- computational intelligence
- artificial intelligence
- active learning
- semi supervised learning
- soft computing
- training samples
- real time
- machine learning
- test set
- training data
- artificial neural networks
- learning tasks
- labeled data
- search algorithm
- pairwise
- training set