ACL RD-TEC 1.0 Summarization of W00-0714
Paper Title:
USING PERFECT SAMPLING IN PARAMETER ESTIMATION OF A WHOLE SENTENCE MAXIMUM ENTROPY LANGUAGE MODEL
USING PERFECT SAMPLING IN PARAMETER ESTIMATION OF A WHOLE SENTENCE MAXIMUM ENTROPY LANGUAGE MODEL
Authors: F. Amayat and J. M. Benedi
Primarily assigned technology terms:
- algorithm
- coupling
- estimation process
- independence metropolis-hastings
- iterative algorithm
- iterative scaling
- language modeling
- learning
- maximum entropy
- maximum entropy framework
- maximum entropy model
- metropolis-hastings
- metropolis-hastings algorithm
- modeling
- normalization
- parameter estimation
- sampling
Other assigned terms:
- approach
- case
- conditional probabilities
- correlation
- distribution
- entropy
- estimation
- events
- feature
- grammatical information
- language model
- language models
- log-likelihood
- markov chain
- maximum entropy principle
- method
- modeling problem
- n-gram
- n-gram model
- n-grams
- parameter values
- perplexity
- prior distribution
- probabilities
- probability
- probability distribution
- process
- random sample
- sentence
- sentences
- structural information
- structure of the sentence
- syntactic information
- technique
- test set
- training
- training corpus
- training set
- training time
- trigram
- trigram model
- vocabulary
- wall street journal corpus
- words
