ASR-Generated Text for Language Model Pre-training Applied to Speech Tasks.
Valentin PelloinFranck DaryNicolas HervéBenoît FavreNathalie CamelinAntoine LaurentLaurent BesacierPublished in: CoRR (2022)
Keyphrases
- language model
- speech recognition
- word error rate
- automatic speech recognition
- language modeling
- information retrieval
- n gram
- speech signal
- query expansion
- document retrieval
- probabilistic model
- error rate
- handwriting recognition
- language modelling
- multiword
- context sensitive
- noisy environments
- hidden markov models
- text retrieval
- retrieval model
- test collection
- smoothing methods
- mixture model
- statistical language models
- vector space model
- translation model
- natural language processing
- language models for information retrieval