Improving Low-Resource Speech Recognition with Pretrained Speech Models: Continued Pretraining vs. Semi-Supervised Training.
Mitchell DeHavenJayadev BillaPublished in: CoRR (2022)
Keyphrases
- speech recognition
- acoustic models
- speech signal
- speech synthesis
- automatic speech recognition
- speech recognizer
- hidden markov models
- language model
- speaker identification
- noisy environments
- probabilistic model
- speech processing
- speech recognition technology
- pattern recognition
- speaker independent
- speech recognition systems
- speaker recognition
- supervised training
- speaker dependent
- speech recognition errors
- noisy speech
- vocal tract
- text to speech
- broadcast news
- neural nets
- generative model
- model selection
- image processing
- information retrieval