ACL RD-TEC 1.0 Summarization of P05-1069
Paper Title:
A LOCALIZED PREDICTION MODEL FOR STATISTICAL MACHINE TRANSLATION
A LOCALIZED PREDICTION MODEL FOR STATISTICAL MACHINE TRANSLATION
Authors: Christoph Tillmann and Tong Zhang
Primarily assigned technology terms:
- algorithm
- approximate inference
- approximation
- arabic-english translation
- beam search
- block alignment
- block translation
- chunking
- classification
- computing
- conditional random field
- convex optimization
- decoder
- decoding
- discriminative re-ranking
- discriminative reranking
- discriminative training
- entropy training
- generalized iterative scaling
- iterative scaling
- language processing
- learning
- lexical weighting
- likelihood training
- logistic regression
- machine translation
- matching
- matching algorithm
- maximum entropy
- maximum likelihood
- maximum likelihood training
- maximum-entropy
- maximum-likelihood
- modeling
- monitoring
- natural language processing
- neural networks
- nlp
- normalization
- online training
- optimization
- parameter training
- parsing
- part-of-speech tagging
- pattern matching
- perceptron
- perceptron algorithm
- perceptron training
- phrase chunking
- phrase translation
- phrase-pair selection
- pos tagging
- post-processing
- processing
- re-ranking
- regression
- regularization
- reranking
- search
- segmentation
- segmentation process
- selection algorithm
- shallow parsing
- smoothing
- statistical machine translation
- statistical parsing
- stochastic optimization
- support vector machines
- tagging
- tokenization
- training algorithm
- training method
- training procedure
- translation process
- translation system
- viterbi
- weighting
- word alignment
Other assigned terms:
- annotation
- approach
- arabic-english translation task
- arabic-to-english translation task
- baseline model
- beam
- bigram
- bigram model
- binary feature
- binary features
- bleu
- case
- concept
- conditional probability
- convergence
- decision rule
- entropy
- entropy formulation
- evaluation test
- events
- experimental results
- fact
- feature
- feature types
- feature vector
- feature weights
- generation
- heuristic
- interpretation
- knowledge
- language annotation
- language model
- language model probability
- language model score
- learning rate
- learning rate parameter
- likelihood
- log-likelihood
- log-linear combination
- log-linear model
- log-linear models
- method
- methodology
- model parameters
- model performance
- model probability
- natural language
- nlp applications
- paragraph
- parallel sentence
- part-of-speech
- phrase
- phrase translation accuracy
- phrase-based model
- possible translation
- prediction model
- probabilities
- probability
- probability model
- procedure
- process
- punctuation
- reference translations
- relative frequency
- search procedure
- sentence
- sentence level
- sentence pair
- sentences
- source sentence
- standard arabic-english translation task
- statistical model
- statistics
- support vector
- syntactic knowledge
- target sentence
- target word
- technique
- term
- terms
- test data
- test set
- theorem
- theory
- training
- training criterion
- training data
- training instance
- translation accuracy
- translation model
- translation probability
- translation score
- translation task
- translations
- trigram
- trigram language model
- unigram
- unigram probability
- weight vector
- weighting scheme
- word
- words
