ACL RD-TEC 1.0 Summarization of H01-1058
Paper Title:
ON COMBINING LANGUAGE MODELS: ORACLE APPROACH
ON COMBINING LANGUAGE MODELS: ORACLE APPROACH
Authors: K. Hacioglu and W. Ward
Primarily assigned technology terms:
- algorithm
- approximation
- communicator
- decision tree
- dialog system
- efficient parsing
- em algorithm
- language modeling
- learning
- linear interpolation
- list rescoring
- maximum entropy
- maximum entropy approach
- modeling
- neural network
- normalization
- optimization
- parser
- parsing
- recognition
- recognizer
- reporting
- rescoring
- search
- sentence generation
- smoothing
- speech recognition
- speech recognition and understanding
- speech recognizer
- statistical language modeling
- tagging
- transcription
- viterbi
Other assigned terms:
- acoustic model
- ambiguity
- approach
- break
- case
- concept
- concepts
- context free grammars
- data sparseness
- dialog
- dialog context
- distribution
- dynamic language model
- entropy
- error rate
- estimation
- experimental results
- feature
- feature vector
- feature vectors
- generation
- grammar
- grammars
- heuristics
- hypothesis
- interpolation
- language model
- language models
- large training
- lexicon
- meaning
- measures
- method
- model combination
- model parameters
- model structure
- n-best list
- n-gram
- n-grams
- normalization factor
- oracle
- parse
- part of speech
- perplexity
- phrase
- phrase level
- portability
- probabilities
- probability
- probability distribution
- production rules
- semantic
- semantic accuracy
- semantic grammar
- semantic tags
- sentence
- sentences
- signal
- speech signal
- symbols
- tags
- term
- terms
- test set
- text
- text corpus
- theory
- training
- training data
- transition network
- tree
- trigram
- understanding
- user
- word
- word sequence
- word sequences
- word string
- word strings
- words
