Generative Adversarial Training Data Adaptation for Very Low-Resource Automatic Speech Recognition.
Kohei MatsuuraMasato MimuraShinsuke SakaiTatsuya KawaharaPublished in: INTERSPEECH (2020)
Keyphrases
- automatic speech recognition
- training data
- speech recognition
- speaker adaptation
- discriminative training
- acoustic models
- conversational speech
- speech signal
- broadcast news
- word error rate
- hidden markov models
- speech retrieval
- generative model
- recognition errors
- noisy environments
- discriminative methods
- spoken words
- decision trees
- spontaneous speech
- word recognition
- learning algorithm
- active learning
- speech corpus
- acoustic features
- training process
- maximum likelihood
- speaker identification
- language model
- pattern recognition