ACL RD-TEC 1.0 Summarization of W04-3007
Paper Title:
ROBUSTNESS ISSUES IN A DATA-DRIVEN SPOKEN LANGUAGE UNDERSTANDING SYSTEM
ROBUSTNESS ISSUES IN A DATA-DRIVEN SPOKEN LANGUAGE UNDERSTANDING SYSTEM
Authors: Yulan He and Steve Young
Primarily assigned technology terms:
- algorithm
- approximation
- atis
- communicator
- continuous speech recognizer
- darpa communicator
- database
- decoder
- decoding
- dialog act decoding
- dialogue act detection
- dialogue systems
- discriminant analysis
- generalized iterative scaling
- hidden markov
- hidden markov model
- iterative scaling
- language model adaptation
- language modelling
- language understanding
- language understanding system
- linear discriminant
- linear discriminant analysis
- linear interpolation
- linear regression
- map adaptation
- markov model
- maximum likelihood
- measuring
- model adaptation
- modelling
- naive bayes
- naive bayes networks
- normalization
- optimisation
- parser
- parsers
- parsing
- push-down automata
- recogniser
- recognition
- recognizer
- regression
- relational database
- robust parsing
- search
- semantic parser
- semantic parsing
- speaker adaptation
- speech recognition
- speech recognizer
- spoken dialogue
- spoken dialogue systems
- spoken language systems
- spoken language understanding
- statistical parser
- statistical parsers
- system development and evaluation
- word generation
Other assigned terms:
- acoustic model
- acoustic signal
- additive noise
- ambiguity
- automata
- background noise
- baseline model
- case
- class information
- concept
- concepts
- conditional probability
- continuous speech
- corpora
- detection rate
- dialog
- dialog act
- dialogue act
- distance optimisation problem
- distribution
- error rate
- estimation
- experimental results
- f-measure
- fact
- frame
- generation
- grammar
- grammar rules
- heuristic
- heuristic rules
- hierarchical structure
- human speech
- hypotheses
- independence assumption
- interpolation
- joint probability
- kullback-leibler distance
- language model
- language models
- lattice
- lexical class
- likelihood
- maps
- meaning
- meanings
- measure
- method
- model parameters
- model probability
- multinomial distribution
- mutual information
- n-best list
- n-gram
- n-gram language model
- nist
- noise
- noisy input
- optimisation problem
- parse
- parser performance
- pcfg
- performance evaluation
- portability
- posterior
- posterior probability
- precision
- prior distribution
- prior probability
- probabilities
- probability
- queries
- query
- recognition accuracy
- recognition component
- recognition errors
- representations
- search space
- semantic
- semantic grammar
- semantic parse
- semantic representation
- semantic representations
- semantic tag
- semantic tags
- sentence
- sentences
- shift operation
- signal
- slot
- speech data
- speech recognition performance
- speech signal
- spoken language
- statistical framework
- statistical model
- style
- system development
- system performance
- tags
- technique
- term
- terms
- test data
- test set
- tokens
- topics
- training
- training and test data
- training corpus
- training data
- understanding
- understanding component
- uniform distribution
- user
- utterance
- vocabulary
- vocabulary size
- word
- word error rate
- word lattice
- word lists
- word sequence
- word string
- words
