End-to-end ASR: from Supervised to Semi-Supervised Learning with Modern Architectures.
Gabriel SynnaeveQiantong XuJacob KahnEdouard GraveTatiana LikhomanenkoVineel PratapAnuroop SriramVitaliy LiptchinskyRonan CollobertPublished in: CoRR (2019)
Keyphrases
- end to end
- semi supervised learning
- semi supervised
- supervised learning
- unsupervised learning
- unlabeled data
- semi supervised classification
- labeled data
- multiple instance
- labeled and unlabeled data
- unlabeled training data
- machine learning
- supervised classifiers
- multiple instance learning
- manifold regularization
- active learning
- semi supervised learning algorithms
- co training
- label propagation
- pairwise
- learning algorithm
- transfer learning
- labeled examples
- label information
- congestion control
- training data
- multi view
- graph construction
- supervised methods
- weakly supervised
- labeled training data
- feature selection
- text classification
- learning models
- graph based semi supervised learning
- object recognition
- data mining
- semi supervised learning methods
- supervised classification
- neural network