ACL RD-TEC 1.0 Summarization of J04-4002
Paper Title:
THE ALIGNMENT TEMPLATE APPROACH TO STATISTICAL MACHINE TRANSLATION
THE ALIGNMENT TEMPLATE APPROACH TO STATISTICAL MACHINE TRANSLATION
Authors: Franz Josef Och and Hermann Ney
Primarily assigned technology terms:
- algorithm
- alignment method
- alignment template model
- alignment template system
- approximation
- automatic evaluation
- beam search
- beam search algorithm
- beam search pruning
- breadth-first search
- categorization
- classification
- computational linguistics
- computer science
- computing
- data estimation
- decoder
- dialogue act classification
- direct optimization
- discriminative training
- dynamic programming
- dynamic programming search
- dynamic-programming-based beam search
- em algorithm
- expectation maximization
- generalized iterative scaling
- graph search
- greedy search
- hidden markov
- hidden markov model
- histogram pruning
- instantiation
- integer programming
- iterative scaling
- language understanding
- learning
- lexicon lookup
- log-linear modeling
- machine translation
- machine translation approach
- machine translation evaluation
- machine translation system
- machine translation systems
- markov model
- maximum approximation
- maximum-entropy
- modeling
- mt system
- mt systems
- natural language understanding
- normalization
- optimal integer programming
- optimization
- parallel training
- parser
- phrasal translation
- phrase alignment
- phrase translation
- phrase-based statistical machine translation
- phrase-based translation
- postprocessing
- preprocessing
- processing
- pruning
- scoring
- search
- search algorithm
- search process
- search pruning
- searching
- segmentation
- segmentation tool
- single reference translation
- source-channel modeling
- speech translation
- spelling
- splitting
- statistical alignment
- statistical approaches
- statistical learning
- statistical machine translation
- statistical mt
- statistical techniques
- statistical translation
- symmetrization
- syntax-based translation
- training algorithm
- training procedure
- translation evaluation
- translation search
- translation system
- translation systems
- tuning
- viterbi
- viterbi alignment
- word alignment
- word selection
Other assigned terms:
- 5-gram language model
- alignment information
- alignment model
- alignment models
- alignment probability
- alignment template
- annotated corpora
- appointment scheduling
- approach
- association for computational linguistics
- automatic evaluation metric
- beam
- beam size
- bigram
- bilingual lexicon
- bilingual phrase
- bleu
- bleu score
- bleu scores
- case
- case information
- characters
- chinese characters
- chinese language
- chinese text
- chinese words
- compound words
- concept
- concepts
- convergence
- corpora
- data consortium
- decision rule
- decision theory
- dialogue act
- dictionary
- edit distance
- english corpus
- english translation
- english vocabulary
- error rate
- estimation
- evaluation metric
- evaluations
- experimental results
- fact
- feature
- french
- french word
- generation
- german-english verbmobil task
- grammatical features
- heuristic
- heuristics
- histogram
- hypotheses
- hypothesis
- implementation
- index
- interpolation
- knowledge
- language model
- language model perplexity
- language model probability
- language models
- language pairs
- language understanding task
- lexica
- lexicon
- lexicon model
- likelihood
- linguistic
- linguistic data
- linguistic data consortium
- linguistics
- log-linear model
- log-linear models
- machine translation model
- machine translation research
- main verb
- meanings
- measure
- measures
- memory consumption
- method
- model parameters
- model perplexity
- model probability
- mt evaluation
- mt quality
- n-best list
- n-gram
- n-grams
- names
- natural language
- natural languages
- nist
- optimization problem
- parallel training corpus
- parts-of-speech
- permutation
- perplexity
- phrase
- phrase level
- polynomial time
- posterior
- posterior probability
- precision
- prior probability
- probabilities
- probability
- probability distributions
- procedure
- process
- reference translation
- reference translations
- reordering
- search problem
- search space
- search time
- search tree
- semantic
- semantic categories
- semantic features
- sentence
- sentence pair
- sentences
- size of the corpus
- source language
- source language sentence
- source language word
- source sentence
- speech translation task
- statistical alignment models
- statistical decision theory
- statistical model
- statistical models
- statistical translation model
- statistics
- subtrees
- symbol
- symbols
- target language
- target language sentence
- target sentence
- target word
- technology
- term
- test corpus
- test data
- text
- theory
- training
- training and test data
- training corpus
- training criterion
- training data
- training material
- translation candidates
- translation hypothesis
- translation lexicon
- translation model
- translation models
- translation probability
- translation quality
- translation research
- translation task
- translations
- tree
- trigram
- trigram language model
- trigram model
- understanding
- unigram
- unigram model
- verb
- vocabulary
- word
- word alignments
- word classes
- word error rate
- word order
- word-aligned corpus
- words
- world knowledge