Minimum Word Error Rate Training with Language Model Fusion for End-to-End Speech Recognition.
Zhong MengYu WuNaoyuki KandaLiang LuXie ChenGuoli YeEric SunJinyu LiYifan GongPublished in: Interspeech (2021)
Keyphrases
- end to end
- speech recognition
- word error rate
- language model
- automatic speech recognition
- language modeling
- handwriting recognition
- n gram
- probabilistic model
- error rate
- document retrieval
- test collection
- query expansion
- hidden markov models
- mixture model
- speech signal
- noisy environments
- information retrieval
- multimedia
- broadcast news
- speaker identification
- retrieval model
- query terms
- pattern recognition