Kaizen: Continuously Improving Teacher Using Exponential Moving Average for Semi-Supervised Speech Recognition.
Vimal ManoharTatiana LikhomanenkoQiantong XuWei-Ning HsuRonan CollobertYatharth SarafGeoffrey ZweigAbdelrahman MohamedPublished in: ASRU (2021)
Keyphrases
- speech recognition
- moving average
- semi supervised
- autoregressive
- hidden markov models
- speech synthesis
- automatic speech recognition
- pattern recognition
- speech recognizer
- language model
- speech signal
- speech processing
- speech recognition systems
- pairwise
- learning process
- isolated word
- speaker identification
- noisy environments
- non stationary
- speaker independent
- spatial domain
- active learning
- feature extraction
- speaker adaptation
- speech recognizers
- information retrieval