ACL RD-TEC 1.0 Summarization of W03-0413
Paper Title:
CONFIDENCE ESTIMATION FOR TRANSLATION PREDICTION
CONFIDENCE ESTIMATION FOR TRANSLATION PREDICTION
Authors: Simona Gandrabur and George Foster
Primarily assigned technology terms:
- algorithm
- beam search
- computing
- confidence estimation
- decision-making
- decoding
- editing
- feature selection
- heuristic search
- ibm translation
- learning
- learning approaches
- learning techniques
- loglinear
- machine learning
- machine learning approaches
- machine learning techniques
- machine translation
- maxent
- maximum entropy
- maximum entropy model
- maximum likelihood
- maximum-entropy
- measuring
- multi-layer perceptron
- neural net
- nlp
- normalization
- perceptron
- post-prediction confidence estimation
- prediction system
- re-scoring
- reading
- scoring
- scoring function
- search
- statistical machine translation
- statistical translation
- tagging
- translator
- translators
- voting
- voting scheme
- word alignment
Other assigned terms:
- approach
- bayes model
- beam
- bias
- case
- character sequence
- characters
- conditional probabilities
- conditional probability
- confidence score
- distribution
- entropy
- entropy models
- estimation
- evaluations
- fact
- feature
- heuristic
- hypotheses
- ibm model
- joint distribution
- language model
- likelihood
- log-likelihood
- maxent model
- maximum entropy models
- meaning
- measure
- model combination
- model probability
- nlp task
- normalization factor
- oracle
- parallel corpus
- posterior
- prediction accuracy
- prediction model
- prediction task
- priori
- probabilities
- probability
- probability estimates
- procedure
- process
- reference translation
- relation
- search procedure
- sentence
- source sentence
- source text
- statistical translation model
- target sentence
- target text
- term
- terms
- text
- the discrimination capacity
- tokens
- training
- training data
- translation component
- translation model
- translation models
- trigram
- trigram language model
- user
- user behavior
- user model
- weight vector
- word
- word sequence
- words
