ACL RD-TEC 1.0 Summarization of P06-1097
Paper Title:
SEMI-SUPERVISED TRAINING FOR STATISTICAL WORD ALIGNMENT
SEMI-SUPERVISED TRAINING FOR STATISTICAL WORD ALIGNMENT
Authors: Alexander Fraser and Daniel Marcu
Primarily assigned technology terms:
- algorithm
- alignment algorithm
- approximation
- classi cation
- classi er
- clustering
- computational linguistics
- computing
- discriminative approach
- discriminative reranking
- discriminative training
- em training
- error minimization
- error reduction
- expectation maximization
- giza
- heuristic symmetrization
- high performance computing
- hillclimbing
- learning
- learning task
- likelihood training
- machine translation
- maximum likelihood
- maximum likelihood training
- measuring
- modeling
- naive bayes
- one-dimensional error minimization
- processing
- reinforcement learning
- reranking
- sampling
- search
- search algorithm
- searching
- semi-supervised algorithm
- semi-supervised learning
- semi-supervised training
- smt system
- statistical machine translation
- statistical word alignment
- supervised learning
- symmetrization
- text classi cation
- tuning
- unsupervised learning
- unsupervised training
- viterbi
- viterbi alignment
- viterbi training
- word aligner
- word alignment
- word-alignment
Other assigned terms:
- approach
- association for computational linguistics
- bleu
- bleu scores
- case
- classi cation task
- clusters
- corpora
- data set
- data sets
- distribution
- estimation
- f-measure
- foreign word
- foreign words
- french
- gold standard
- heuristic
- heuristics
- hmm model
- hypotheses
- hypothesis
- ibm model
- ibm models
- knowledge
- language models
- language pairs
- large training
- large training corpora
- likelihood
- linguistics
- log-linear model
- machine translation performance
- meaning
- measure
- mt quality
- n-best list
- precision
- probability
- probability distribution
- procedure
- process
- relation
- search problem
- semi-supervised approach
- sentence
- sentence pair
- sentences
- source language
- stem
- supervised model
- technique
- test set
- text
- training
- training corpora
- training corpus
- training data
- training set
- translation quality
- translation table
- translation test set
- translations
- understanding
- vocabulary
- word
- word alignments
- words
