From Semi-supervised to Almost-unsupervised Speech Recognition with Very-low Resource by Jointly Learning Phonetic Structures from Audio and Text Embeddings.
Yi-Chen ChenSung-Feng HuangHung-yi LeeLin-Shan LeePublished in: CoRR (2019)
Keyphrases
- speech recognition
- semi supervised
- supervised learning
- unsupervised learning
- speech processing
- hidden markov models
- speech recognition technology
- speaker identification
- active learning
- pattern recognition
- language model
- speech recognizer
- learning process
- noisy environments
- labeled data
- speech signal
- automatic speech recognition
- speech synthesis
- text classification
- multimedia
- machine learning