ACL RD-TEC 1.0 Summarization of N06-1013
Paper Title:
A MAXIMUM ENTROPY APPROACH TO COMBINING WORD ALIGNMENTS
A MAXIMUM ENTROPY APPROACH TO COMBINING WORD ALIGNMENTS
Authors: Necip Fazil Ayan and Bonnie J. Dorr
Primarily assigned technology terms:
- algorithm
- alignment combination
- alignment process
- bilingual sense disambiguation
- bootstrap
- bootstrap resampling
- classification
- classifier
- classifier ensemble
- computing
- cross validation
- crosslanguage information retrieval
- decoding
- disambiguation
- error reduction
- exhaustive search
- extractor
- framenet classification
- generalized iterative scaling
- giza
- heuristic combination
- information retrieval
- iterative scaling
- kneser-ney smoothing
- language modeling
- learning
- learning algorithm
- learning approach
- learning method
- machine translation
- matching
- maximum entropy
- maximum entropy approach
- maximum entropy model
- modeling
- mt system
- neural network
- nlp
- normalization
- parsing
- partitioning
- pos tagger
- pos tagging
- ranking
- resampling
- search
- segmentation
- sense disambiguation
- sentence segmentation
- smoothing
- sri language modeling
- statistical alignment
- statistical classification
- statistical machine translation
- supervised learning
- supervised learning approach
- tagger
- tagging
- unsupervised learning
- validation
- word alignment
- word reordering
- word-alignment
- word-sense disambiguation
Other assigned terms:
- aligned corpus
- alignment error rate
- alignment problem
- approach
- bilingual lexicons
- bleu
- bleu scores
- brevity penalty
- case
- chinese treebank
- classification problem
- confidence score
- data set
- distribution
- document
- english sentence
- entropy
- entropy models
- error rate
- evaluation data
- evaluation metric
- evaluation metrics
- evaluations
- f-score
- feature
- feature set
- feature sets
- framenet
- gold standard
- heuristic
- heuristics
- ibm model
- ibm models
- implementation
- knowledge
- labeling
- language model
- language modeling toolkit
- language pair
- language pairs
- likelihood
- linguistic
- linguistic features
- log-linear combination
- maximum entropy models
- method
- model parameter
- model parameters
- modeling toolkit
- mt evaluation
- mt quality
- n-grams
- neural network framework
- nist
- oracle
- parallel corpus
- parallel texts
- part-of-speech
- part-of-speech tags
- penn treebank
- phrase
- pp attachment
- precision
- probability
- probability distribution
- process
- projection
- relative error reduction
- reordering
- sentence
- sentence level
- sentences
- tags
- test material
- test set
- text
- toolkit
- training
- training data
- training set
- translation model
- translational equivalence
- translations
- treebank
- trigram
- trigram model
- wall street journal corpus
- word
- word alignments
- words
- yasmet me package
