A robust language model incorporating a substring parser and extended n-grams
J. H. Wright, Gareth J. F. Jones, Harvey Lloyd-Thomas · IEEE International Conference on Acoustics Speech and Signal Processing · 2002
Describes a language model for speech recognition which incorporates a substring parser (to take advantage of syntactic structure covered by a context-free grammar) and extended bigrams (to take advantage of remote dependencies between words). The use of extended bigrams significantly reduces the perplexity and a distribution clustering algorithm alleviates the additional storage cost. The substring parser is the foundation for training and scoring procedures based on paths at all levels through the syntactic structures, with subtrees linked by bigrams. The word bigram score is therefore absorbed into a grammar framework, consolidating the two kinds of language model, and again a significant reduction in perplexity is observed. The aim is an integrated, robust language model that is adaptive to the speaker.>
Cited by 9 works. doi:10.1109/icassp.1994.389281
Authors
References
- The SPHINX-II speech recognition system: an overview
- Efficient Parsing for Natural Language
- Trigger-based language models: a maximum entropy approach
- Improvements in stochastic language modeling
- The ARM continuous speech recognition system
- Statistical language modeling combining N-gram and context-free grammars
- Language modelling for recognition and understanding using layered bigrams
- A consolidated language model for speech recognition
- The HMM interface with hybrid grammar-bigram language models for speech recognition
- Computational requirements of probabilistic LR parsing for speech recognition using a natural language grammar