ACL RD-TEC 1.0 Summarization of N04-4018
Paper Title:
IMPROVING AUTOMATIC SENTENCE BOUNDARY DETECTION WITH CONFUSION NETWORKS
IMPROVING AUTOMATIC SENTENCE BOUNDARY DETECTION WITH CONFUSION NETWORKS
Authors: D. Hillard and M. Ostendorf and A. Stolcke and Y. Liu and E. Shriberg
Primarily assigned technology terms:
- automatic speech recognition
- bagging
- boundary classification
- boundary detection
- boundary prediction
- classification
- classifier
- decision tree
- decision tree classifier
- decision tree training
- decoding
- disfluency detection
- evaluation tools
- feature combination
- feature extraction
- feature normalization
- finitestate
- language modeling
- language processing
- maximum entropy
- maximum entropy model
- modeling
- neural net
- normalization
- parsing
- processing
- prosodic feature normalization
- recognition
- recognition system
- recognition systems
- recognizer
- rescoring
- scoring
- scoring tool
- segmentation
- sentence boundary detection
- speech recognition
- speech recognition system
- speech recognition systems
- sri language modeling
- su detection
- summarization
- topic segmentation
- transcription
- tree classifier
- word recognition
Other assigned terms:
- alignment procedure
- annotation
- approach
- broadcast news
- case
- classification tree
- community
- confidence score
- confusion network
- conversation
- conversational telephone speech
- corpora
- development set
- duration
- entropy
- error rate
- evaluation data
- evaluation set
- evaluation test
- evaluations
- events
- experimental results
- fact
- feature
- human understanding
- hypotheses
- hypothesis
- interpolation
- knowledge
- language model
- language modeling toolkit
- language models
- lattice
- lattices
- lexical feature
- linear combination
- main clause
- metadata
- method
- modeling toolkit
- n-best list
- nist
- opinions
- oracle
- pause
- pause information
- pos-based model
- posterior
- posterior probability
- precision
- probabilities
- probability
- procedure
- pronunciation
- prosodic feature
- prosodic features
- prosodic information
- prosody
- punctuation
- recognition errors
- representations
- segments
- sentence
- sentence boundaries
- sentence boundary
- sentence structure
- sentences
- slot
- speaker change
- speaker change information
- speech recognition errors
- statistics
- switchboard corpus
- terms
- test data
- tokens
- toolkit
- topology
- training
- training data
- transcripts
- tree
- trees
- understanding
- word
- word boundaries
- word boundary
- word classes
- word confusion network
- word error rate
- word error rates
- word lattices
- word recognition output
- word sequence
- word strings
- words
