ACL RD-TEC 1.0 Summarization of W06-2601
Paper Title:
MAXIMUM ENTROPY TAGGING WITH BINARY AND REAL-VALUED FEATURES
MAXIMUM ENTROPY TAGGING WITH BINARY AND REAL-VALUED FEATURES
Authors: Vanessa Sandrini and Marcello Federico and Mauro Cettolo
Primarily assigned technology terms:
- algorithm
- capitalization
- chunking
- classification
- discounting method
- dynamic programming
- entity recognition
- generalized iterative scaling
- iterative scaling
- language modeling
- leave-one-out method
- machine translation
- maximum entropy
- maximum likelihood
- me training
- modeling
- named entity recognition
- named-entity recognition
- nlp
- parameter estimation
- recognition
- sampling
- smoothing
- smoothing techniques
- statistical language modeling
- statistical machine translation
- taggers
- tagging
- text chunking
- text-tagging
- training algorithm
Other assigned terms:
- binary feature
- binary features
- capitalization information
- chunk
- classification task
- co-occurrences
- conditional distribution
- conditional probabilities
- conditional probability
- context size
- correlations
- data sparseness
- dictionary
- discriminative model
- distribution
- entropy
- estimation
- events
- experimental results
- fact
- feature
- feature type
- formalism
- index
- interpolation
- lexical feature
- lexical features
- likelihood
- log-likelihood
- measure
- measures
- method
- model complexity
- model parameters
- named entity
- named-entity
- names
- nlp tasks
- part-of-speech
- posterior
- posterior probability
- probabilities
- probability
- probability distribution
- probability distributions
- proper names
- statistical framework
- statistics
- symbol
- syntactic features
- syntactic information
- tags
- term
- terms
- test set
- text
- theory
- training
- training corpus
- training data
- training time
- word
- words
