Accent-Specific Vector Quantization for Joint Unsupervised and Supervised Training in Accent Robust Speech Recognition.
Li LiYijie LiDongxing XuHaoran WeiYanhua LongPublished in: ICASSP (2024)
Keyphrases
- speech recognition
- vector quantization
- supervised training
- image compression
- noisy environments
- hidden markov models
- vector quantizer
- automatic speech recognition
- unsupervised learning
- conditional random fields
- speaker recognition
- speech signal
- supervised learning
- training phase
- pattern recognition
- language model
- speaker identification
- recognition process
- semi supervised
- dependency parsing
- signal to noise ratio
- broadcast news
- bayesian networks