Developing a hybrid language model for open vocabulary automatic speech recognition in a lecture speech task.
Marc-Antoine RondeauRichard C. RosePublished in: ISSPA (2012)
Keyphrases
- automatic speech recognition
- language model
- speech recognition
- word error rate
- spoken term detection
- speech signal
- language modeling
- out of vocabulary
- broadcast news
- speech synthesis
- information retrieval
- document retrieval
- n gram
- query expansion
- retrieval model
- test collection
- probabilistic model
- error rate
- recognition errors
- noisy environments
- speech retrieval
- language modelling
- conversational speech
- speech corpus
- word recognition
- spontaneous speech
- hidden markov models
- search engine
- smoothing methods
- acoustic features
- handwriting recognition
- vector space model
- query terms
- mixture model
- keywords
- term dependencies
- spoken document retrieval
- relevance model
- pseudo relevance feedback
- speaker adaptation