Using Semi-supervised Learning for Monaural Time-domain Speech Separation with a Self-supervised Learning-based SI-SNR Estimator.
Shaoxiang DangTetsuya MatsumotoYoshinori TakeuchiHiroaki KudoPublished in: INTERSPEECH (2023)
Keyphrases
- semi supervised learning
- supervised learning
- learning problems
- unlabeled data
- active learning
- labeled and unlabeled data
- semi supervised
- unsupervised learning
- labeled data
- graph transduction
- co training
- learning algorithm
- learning process
- prior knowledge
- multi view learning
- learning models
- domain adaptation
- supervised learning algorithms
- covariate shift
- data sets
- learning tasks
- training data
- labeled examples
- boosting framework
- graph based semi supervised learning
- kernel methods
- semi supervised classification
- incremental learning
- maximum likelihood
- information extraction
- pattern recognition
- machine learning