Efficient minimum word error rate training of RNN-Transducer for end-to-end speech recognition.
Jinxi GuoGautam TiwariJasha DroppoMaarten Van SegbroeckChe-Wei HuangAndreas StolckeRoland MaasPublished in: CoRR (2020)
Keyphrases
- speech recognition
- end to end
- word error rate
- automatic speech recognition
- language model
- speech signal
- isolated word
- handwriting recognition
- error rate
- noisy environments
- speech synthesis
- speech recognizer
- hidden markov models
- recurrent neural networks
- pattern recognition
- speaker identification
- acoustic models
- speech recognition systems
- language modeling
- speaker independent
- machine learning
- broadcast news
- face recognition
- noise reduction