ACL RD-TEC 1.0 Summarization of N03-1019
Paper Title:
A WEIGHTED FINITE STATE TRANSDUCER IMPLEMENTATION OF THE ALIGNMENT TEMPLATE MODEL FOR STATISTICAL MACHINE TRANSLATION
A WEIGHTED FINITE STATE TRANSDUCER IMPLEMENTATION OF THE ALIGNMENT TEMPLATE MODEL FOR STATISTICAL MACHINE TRANSLATION
Authors: Shankar Kumar and William Byrne
Primarily assigned technology terms:
- algorithm
- alignment process
- alignment template model
- bitext alignment
- bitext word alignment
- call routing
- decoder
- dynamic programming
- finite state
- finite state composition
- finite state modeling
- finite state transducer
- finite state transducers
- fsm toolkit
- giza
- hansards translation
- kneser-ney smoothing
- lexicon transducer
- machine translation
- maximum likelihood
- modeling
- normalization
- parameter estimation
- phrasal translation
- phrase alignment
- phrase segmentation
- pruning
- search
- segmentation
- smoothing
- statistical machine translation
- transducer
- transducers
- translation process
- tuning
- weighted finite state acceptor
- wfst modeling
- word alignment
- word-to-word alignment
- word-to-word translation
Other assigned terms:
- alignment error rate
- alignment task
- alignment template
- approach
- backoff
- bitext
- bleu
- bleu score
- bleu scores
- community
- composition
- conditional independence
- derivation
- dictionary
- distribution
- error rate
- estimation
- french
- french sentence
- generation
- hypotheses
- ibm model
- implementation
- index
- joint distribution
- joint probability
- joint probability distribution
- language model
- lattice
- lattices
- lexicon
- likelihood
- mapping
- maps
- markov chain
- measure
- model parameters
- n-best list
- nist
- permutation
- phrase
- precision
- probabilities
- probability
- probability distribution
- procedure
- process
- reference translations
- relative frequency
- reordering
- search procedure
- segments
- semiring
- sentence
- sentence pair
- sentences
- sequence model
- source language
- source language sentence
- source sentence
- symbol
- tags
- target language
- target language model
- target language sentence
- target sentence
- target word
- template translation model
- term
- terms
- test data
- test set
- tokens
- toolkit
- training
- training and test data
- training corpus
- training set
- translation dictionary
- translation direction
- translation model
- translation models
- translation quality
- translation task
- translations
- trigram
- trigram language model
- unigram
- vocabulary
- word
- word alignments
- word classes
- word error rate
- word order
- word sequences
- words
