ACL RD-TEC 1.0 Summarization of W06-2606
Paper Title:
RERANKING TRANSLATION HYPOTHESES USING STRUCTURAL PROPERTIES
RERANKING TRANSLATION HYPOTHESES USING STRUCTURAL PROPERTIES
Authors: Saşa Hasan and Oliver Bender and Hermann Ney
Primarily assigned technology terms:
- algorithm
- analyzer
- chunk parser
- chunker
- chunking
- count-based feature reduction
- decoder
- dependency analysis
- dependency analyzer
- direct modeling
- downhill simplex algorithm
- entity recognition
- generalized iterative scaling
- grammar parser
- graph rescoring
- iterative scaling
- language understanding
- lexicalization
- linear interpolation
- link grammar
- link grammar parser
- linking
- machine translation
- machine translation system
- matching
- maximum entropy
- maximum-entropy
- me chunking
- modeling
- named entity recognition
- natural language understanding
- normalization
- parser
- parsing
- part-of-speech tagging
- phrase-based decoder
- phrase-based translation
- recognition
- reranking
- rescoring
- search
- search process
- shallow parsing
- simplex algorithm
- smoothing
- smoothing method
- smt system
- statistical machine translation
- statistical machine translation system
- subcategorization
- supertagging
- tagger
- tagging
- text chunking
- training procedure
- translation system
- tuning
Other assigned terms:
- adjunction
- anchor
- anchors
- annotated corpus
- approach
- bayes decision rule
- bleu
- bleu score
- case
- chunk
- chunk tag
- chunks
- compact representation
- computational complexity
- concept
- corpora
- decision rule
- development set
- dictionary
- elementary tree
- entropy
- error rate
- feature
- formalism
- gaussian prior
- grammar
- grammar formalism
- grammars
- hypotheses
- hypothesis
- ibm model
- interpolation
- language model
- language model score
- language models
- language pairs
- lexical item
- lexicalized tree
- lexicon
- linear time
- local context
- log-linear model
- ltags
- matching process
- maximum entropy principle
- measure
- method
- model parameters
- multilingual speech
- n-best list
- n-gram
- n-gram models
- n-grams
- named entity
- natural language
- nist
- normalization factor
- oracle
- parse
- parsing process
- part-of-speech
- phrase
- pos tag
- posterior
- posterior probability
- probability
- procedure
- process
- reference translation
- reordering
- right-hand side
- search space
- sentence
- sentences
- shallow parse
- source sentence
- speech corpus
- statistical framework
- statistics
- subcategorization frames
- supertag
- supertags
- syntactic feature
- syntactic features
- syntactic structure
- tags
- target language
- target language model
- technique
- terms
- test data
- test set
- text
- theoretical framework
- training
- training corpus
- translation model
- translation models
- translation quality
- translations
- travel expression corpus
- tree
- tree adjoining grammar
- tree structures
- trees
- trigram
- understanding
- word
- word classes
- word error rate
- words
- wsj corpus
- xtag
