VioLA: Unified Codec Language Models for Speech Recognition, Synthesis, and Translation.
Tianrui WangLong ZhouZiqiang ZhangYu WuShujie LiuYashesh GaurZhuo ChenJinyu LiFuru WeiPublished in: CoRR (2023)
Keyphrases
- speech recognition
- language model
- translation model
- statistical machine translation
- language modeling
- document retrieval
- out of vocabulary
- information retrieval
- probabilistic model
- n gram
- query expansion
- automatic speech recognition
- speech signal
- mixture model
- retrieval model
- speech recognition systems
- query terms
- test collection
- speech synthesis
- speech recognizer
- cross language
- query translation
- cross language information retrieval
- speaker identification
- handwriting recognition
- noisy environments
- machine translation
- smoothing methods
- word error rate
- relevance model
- cross lingual
- machine translation system
- pattern recognition
- text categorization
- text mining