DNN-Based Speech Synthesis: Importance of Input Features and Training Data.
Alexandros LazaridisBlaise PotardPhilip N. GarnerPublished in: SPECOM (2015)
Keyphrases
- input features
- speech synthesis
- training data
- feature values
- class labels
- training process
- speech recognition
- support vector machine
- text to speech
- machine learning algorithms
- labeled examples
- learning algorithm
- training set
- classification algorithm
- feature selection
- labeled data
- data sets
- supervised learning
- training samples
- training examples
- classification accuracy
- decision trees
- test data
- unlabeled data
- pattern recognition
- semi supervised learning
- selected features
- feature subset
- prior knowledge
- activation function
- feature space
- basis functions
- pairwise
- bayesian networks
- information retrieval