ACL RD-TEC 1.0 Summarization of P00-1073
Paper Title:
DISTRIBUTION-BASED PRUNING OF BACKOFF LANGUAGE MODELS
DISTRIBUTION-BASED PRUNING OF BACKOFF LANGUAGE MODELS
Authors: Jianfeng Gao and Kai-Fu Lee
Primarily assigned technology terms:
- algorithm
- approximation
- backoff bigram
- backoff model pruning
- bigram distribution modelling
- bigram pruning
- character conversion
- classifier
- clustering
- clustering algorithm
- cutoff
- cutoff method
- cutoff pruning
- cutoff pruning method
- distribution modelling
- distribution-based pruning
- distribution-based pruning method
- information retrieval
- language modelling
- language understanding
- lm pruning
- model pruning
- modelling
- n-gram backoff
- normalization
- pruning
- pruning method
- ranking
- recognition
- speech recognition
- spoken language understanding
- statistical language modelling
- svm classifier
- term distribution modelling
- thresholding
- weighting
Other assigned terms:
- approach
- backoff
- backoff model
- bias
- bigram
- bigram model
- case
- characters
- cluster
- clusters
- conditional probabilities
- content words
- data sparseness
- distribution
- document
- document frequency
- entropy
- estimation
- experimental results
- fact
- finite set
- geometric mean
- implementation
- inverse document frequency
- language model
- language models
- large corpus
- large training
- likelihood
- measure
- method
- mood
- n-gram
- n-gram language model
- n-gram models
- n-grams
- normalization factor
- perplexity
- perplexity reduction
- pinyin
- poisson distribution
- probabilistic model
- probabilities
- probability
- probability estimate
- spoken language
- style
- term
- term distribution
- testing data
- text
- training
- training corpus
- training data
- training set
- training text
- trigram
- understanding
- unigram
- word
- word pair
- word perplexity
- words
