SALM: Speech-augmented Language Model with In-context Learning for Speech Recognition and Translation.
Zhehuai ChenHe HuangAndrei AndrusenkoOleksii HrinchukKrishna C. PuvvadaJason LiSubhankar GhoshJagadeesh BalamBoris GinsburgPublished in: CoRR (2023)
Keyphrases
- speech recognition
- language model
- speech signal
- speech processing
- language modeling
- speech synthesis
- automatic speech recognition
- hidden markov models
- word error rate
- speech recognition systems
- n gram
- speech recognizer
- context sensitive
- probabilistic model
- speaker identification
- pattern recognition
- handwriting recognition
- translation model
- query expansion
- speech recognizers
- noisy environments
- statistical machine translation
- speech recognition technology
- information retrieval
- query terms
- test collection
- unsupervised learning
- speaker diarization
- text mining
- speaker dependent