Virtuoso: Massive Multilingual Speech-Text Joint Semi-Supervised Learning for Text-To-Speech.
Takaaki SaekiHeiga ZenZhehuai ChenNobuyuki MoriokaGary WangYu ZhangAnkur BapnaAndrew RosenbergBhuvana RamabhadranPublished in: CoRR (2022)
Keyphrases
- text to speech
- semi supervised learning
- text to speech synthesis
- speech synthesis
- semi supervised
- unlabeled data
- labeled data
- prosodic features
- machine learning
- supervised learning
- english text
- semi supervised classification
- unsupervised learning
- manifold regularization
- training data
- data analysis
- active learning
- label propagation
- text categorization
- semi supervised learning algorithms
- co training
- word processing
- graph construction
- transfer learning
- labeled and unlabeled data
- text classification
- domain adaptation
- cross lingual
- graph based semi supervised learning
- probabilistic model
- pairwise
- reinforcement learning
- feature extraction