ComSL: A Composite Speech-Language Model for End-to-End Speech-to-Text Translation.
Chenyang LeYao QianLong ZhouShujie LiuMichael ZengXuedong HuangPublished in: CoRR (2023)
Keyphrases
- end to end
- language model
- speech recognition
- translation model
- word error rate
- cross language retrieval
- out of vocabulary
- language modeling
- statistical machine translation
- speech signal
- n gram
- probabilistic model
- language modelling
- information retrieval
- test collection
- document retrieval
- automatic speech recognition
- retrieval model
- query expansion
- machine translation
- query terms
- ad hoc information retrieval
- error rate
- audio visual
- statistical language models
- mixture model
- language models for information retrieval
- cross language information retrieval
- congestion control
- query translation
- context sensitive
- pseudo relevance feedback
- language model for information retrieval
- bilingual dictionaries
- handwriting recognition
- statistical model
- word clouds