ACL RD-TEC 1.0 Summarization of H93-1021
Paper Title:
ADAPTIVE LANGUAGE MODELING USING THE MAXIMUM ENTROPY PRINCIPLE
ADAPTIVE LANGUAGE MODELING USING THE MAXIMUM ENTROPY PRINCIPLE
Authors: Raymond Lau and Ronald Rosenfeld and Salim Roukos
Primarily assigned technology terms:
- adaptive language modeling
- algorithm
- generalized iterative scaling
- hill climbing
- hillclimbing
- incremental adaptation
- iterative algorithm
- iterative scaling
- language modeling
- likelihood estimation
- maximum entropy
- maximum entropy model
- maximum likelihood
- maximum likelihood estimation
- modeling
- probability function
- recognizer
- search
- speech recognizer
- statistical language modeling
Other assigned terms:
- approach
- bigram
- cache
- case
- conditional model
- conditional probabilities
- convergence
- correlations
- distribution
- document
- entropy
- entropy models
- error rate
- estimation
- events
- formalism
- interpolation
- interpretation
- joint distribution
- joint probability
- knowledge
- language model
- language models
- likelihood
- maximum entropy models
- maximum entropy principle
- method
- n-gram
- n-gram model
- perplexity
- perplexity reduction
- probabilities
- probability
- probability distribution
- probability estimate
- probability estimates
- segments
- style
- sublanguage
- term
- test set
- text
- training
- training data
- training set
- trigram
- trigram model
- unigram
- unigram model
- vocabulary
- word
- word sequence
- words
