End-to-end Whispered Speech Recognition with Frequency-weighted Approaches and Layer-wise Transfer Learning.
Heng-Jui ChangAlexander H. LiuHung-yi LeeLin-Shan LeePublished in: CoRR (2020)
Keyphrases
- end to end
- speech recognition
- transfer learning
- application layer
- speech recognition systems
- hidden markov models
- language model
- pattern recognition
- speech signal
- speaker identification
- congestion control
- automatic speech recognition
- knowledge transfer
- machine learning algorithms
- cross domain
- semi supervised learning
- speech synthesis
- reinforcement learning
- neural network
- unlabeled data
- labeled data
- transfer knowledge
- speaker independent
- cepstral coefficients
- semi supervised
- active learning
- pairwise
- learning algorithm
- machine learning
- cross lingual
- target domain
- text categorization
- collaborative filtering
- learning process
- information retrieval