ACL RD-TEC 1.0 Summarization of P04-1064
Paper Title:
ALIGNING WORDS USING MATRIX FACTORISATION
ALIGNING WORDS USING MATRIX FACTORISATION
Authors: Cyril Goutte and Kenji Yamada and Eric Gaussier
Primarily assigned technology terms:
- algorithm
- bilingual terminology
- bilingual terminology extraction
- decomposition
- em algorithm
- expectation-maximisation
- factorisation
- giza
- latent semantic analysis
- learning
- linguistic processing
- machine learning
- machine translation
- machine translation system
- matrix factorisation
- model selection
- modelling
- nolem
- orthogonal non-negative matrix factorisation
- phrase-based machine translation
- phrase-based translation
- phrase-based translation system
- processing
- re-estimation
- semantic analysis
- terminology
- terminology extraction
- thresholding
- translation system
- weighting
- word alignment
- word-based translation
- word-to-word alignment
Other assigned terms:
- aligned sentence
- alignment error rate
- alignment task
- ambiguity
- annotators
- approach
- association measure
- bias
- case
- chunks
- co-occurrence
- concept
- conditional probability
- convergence
- dictionary
- distribution
- english ibm4 models
- error rate
- f-score
- fact
- feature
- french
- function words
- ibm model
- ibm models
- joint probability
- latent semantic
- lexicon
- likelihood
- linguistic
- linguistic features
- log-likelihood
- measure
- measures
- method
- mixture models
- model complexity
- model size
- mutual information
- n-best list
- noise
- orthogonal matrix
- parallel texts
- precision
- probabilistic model
- probabilities
- probability
- process
- projection
- semantic
- sentence
- sentences
- source sentence
- statistical model
- statistics
- target sentence
- technique
- term
- test data
- test set
- training
- transitive closure
- translation models
- translation table
- translations
- uniform distribution
- word
- word alignments
- words
