Transfer Learning, Style Control, and Speaker Reconstruction Loss for Zero-Shot Multilingual Multi-Speaker Text-to-Speech on Low-Resource Languages.
Kurniawati AzizahWisnu JatmikoPublished in: IEEE Access (2022)
Keyphrases
- transfer learning
- prosodic features
- text to speech
- cross lingual
- speaker verification
- speech synthesis
- learning tasks
- knowledge transfer
- prediction with expert advice
- language independent
- active learning
- speech recognition
- learning algorithm
- audio visual
- collaborative filtering
- cross domain
- reinforcement learning
- cross language
- domain adaptation
- automatic speech recognition
- machine learning algorithms
- semi supervised learning
- labeled data
- manifold alignment
- spontaneous speech
- machine learning
- multi task
- machine translation
- target domain
- text categorization
- training data
- information retrieval
- data mining
- text to speech synthesis