Sequence-Level Consistency Training for Semi-Supervised End-to-End Automatic Speech Recognition.
Ryo MasumuraMana IhoriAkihiko TakashimaTakafumi MoriyaAtsushi AndoYusuke ShinoharaPublished in: ICASSP (2020)
Keyphrases
- end to end
- automatic speech recognition
- semi supervised
- speech recognition
- speech signal
- conversational speech
- speech retrieval
- admission control
- high bandwidth
- scalable video
- congestion control
- broadcast news
- acoustic models
- hidden markov models
- word recognition
- pairwise
- word error rate
- recognition errors
- labeled data
- active learning
- spoken words
- video data
- internet protocol
- speech corpus
- bayesian networks