Efficient Implementation of the Room Simulator for Training Deep Neural Network Acoustic Models.
Chanwoo KimEhsan VarianiArun NarayananMichiel BacchianiPublished in: INTERSPEECH (2018)
Keyphrases
- efficient implementation
- neural network
- acoustic models
- training process
- speech recognition
- hidden markov models
- active set
- training set
- speech recognizer
- discriminative training
- broadcast news
- automatic speech recognition
- pattern recognition
- image processing
- back propagation
- speaker identification
- multi modal
- knn
- feature space
- feature extraction
- machine learning