Transfer Learning From Speech Synthesis to Voice Conversion With Non-Parallel Training Data.
Mingyang ZhangYi ZhouLi ZhaoHaizhou LiPublished in: IEEE ACM Trans. Audio Speech Lang. Process. (2021)
Keyphrases
- transfer learning
- speech synthesis
- training data
- text to speech
- labeled data
- target domain
- domain adaptation
- speech recognition
- unlabeled data
- learning tasks
- semi supervised learning
- knowledge transfer
- learning algorithm
- prosodic features
- source domain
- training and test data
- cross domain
- active learning
- vocal tract
- reinforcement learning
- text classification
- supervised learning
- decision trees
- semi supervised
- machine learning algorithms
- text categorization
- manifold alignment
- training set
- multi task
- collaborative filtering
- test set
- label information
- transfer knowledge
- labeled training data
- transferring knowledge
- test data
- prior knowledge
- machine learning
- text mining
- support vector machine
- pattern recognition
- co training
- class labels
- labeled and unlabeled data
- feature space
- bayesian networks
- data mining
- neural network
- data sets