ACL RD-TEC 1.0 Summarization of P02-1025
Paper Title:
A STUDY ON RICHER SYNTACTIC DEPENDENCIES FOR STRUCTURED LANGUAGE MODELING
A STUDY ON RICHER SYNTACTIC DEPENDENCIES FOR STRUCTURED LANGUAGE MODELING
Authors: Peng Xu and Ciprian Chelba and Frederick Jelinek
Primarily assigned technology terms:
- algorithm
- beam search
- binary branching
- bottom-up parser
- classification
- decoding
- em training
- entropy estimation
- language modeling
- left-to-right parsing
- maxent
- maximum entropy
- modeling
- multi-stack search
- multi-stack search algorithm
- n-best rescoring
- normalization
- parser
- parsers
- parsing
- pruning
- re-estimation
- re-estimation algorithm
- re-scoring
- recognition
- recognizer
- rescoring
- right-branching
- rule-based approach
- scoring
- search
- search algorithm
- slm training
- smoothing
- speech recognition
- speech recognizer
- statistical parser
- statistical parsing
- tagger
- training algorithm
- training procedure
- word prediction
Other assigned terms:
- acoustic signal
- annotation
- annotator
- approach
- baseline model
- beam
- case
- community
- complete parse
- contextual information
- corpora
- correlation
- data sparseness
- data sparseness problem
- dependency structure
- dependency structures
- entropy
- estimation
- fact
- human annotator
- hypotheses
- hypothesis
- index
- interpolation
- joint probability
- language model
- language model performance
- language model probability
- lattice
- lattices
- likelihood
- linguistics
- maxent model
- measures
- method
- model parameters
- model performance
- model probability
- nist
- noun phrase
- parse
- parse tree
- parsing accuracy
- part of speech
- perplexity
- phrase
- pos tag
- precision
- predictive annotation
- probabilistic model
- probabilities
- probability
- procedure
- right-hand side
- search space
- search strategy
- sentence
- sentence level
- sentence position
- sentences
- signal
- sparseness problem
- statistics
- structured language model
- symbol
- syntactic structure
- tag information
- tags
- technique
- terminals
- terms
- test data
- test set
- text
- training
- training corpus
- training data
- tree
- treebank
- treebank corpus
- trees
- trigram
- trigram language model
- understanding
- upenn treebank
- utterance
- vocabulary
- vocabulary size
- word
- word level
- word sequence
- word string
- words
- wsj corpora
