Semi-supervised acoustic and language model training for English-isiZulu code-switched speech recognition.
Astik BiswasFebe de WetEwald van der WesthuizenThomas R. NieslerPublished in: CoRR (2020)
Keyphrases
- speech recognition
- language model
- acoustic models
- semi supervised
- speech recognition technology
- speech recognition systems
- isolated word
- speaker independent
- language modeling
- speech recognizers
- speech recognizer
- automatic speech recognition
- n gram
- probabilistic model
- document retrieval
- statistical machine translation
- supervised learning
- information retrieval
- speech synthesis
- semi supervised learning
- speech signal
- retrieval model
- hidden markov models
- query expansion
- word error rate
- cross lingual
- noisy environments
- handwriting recognition
- test collection
- training set
- active learning
- broadcast news
- speaker identification
- mixture model
- query terms
- labeled data
- translation model
- speech retrieval
- natural language
- generative model
- feature vectors
- unlabeled data
- image processing