ACL RD-TEC 1.0 Summarization of H89-1010
Paper Title:
THE BBN BYBLOS CONTINUOUS SPEECH RECOGNITION SYSTEM
THE BBN BYBLOS CONTINUOUS SPEECH RECOGNITION SYSTEM
Authors: Richard Schwartz and Chris Barry and Yen-Lu Chow and Alan Deft and Ming-Whei Feng and Owen Kimball and Francis Kubala and John Makhoul and Jeffrey Vandegrift
Primarily assigned technology terms:
- advanced speech recognition
- algorithm
- beam search
- computing
- continuous speech recognition
- cooccurrence smoothing
- database
- decoding
- forward-backward training
- hidden markov
- hidden markov models
- language understanding
- markov model
- natural language understanding
- recognition
- recognition algorithm
- recognition system
- resource management
- search
- sentence recognition
- smoothing
- smoothing method
- speaker-dependent training
- speaker-independent training
- speech recognition
- speech recognition system
- speech understanding
- spoken language systems
- triphone cooccurrence smoothing
- vector quantization
- viterbi
- viterbi algorithm
Other assigned terms:
- acoustic models
- beam
- byblos system
- codebook
- conditional probability
- context-free grammars
- continuous speech
- density function
- dimensionality
- distribution
- duration
- error rate
- estimation
- evaluation test
- fact
- feature
- feature space
- frame
- grammar
- grammars
- index
- likelihood
- likelihood function
- markov models
- measure
- method
- model parameters
- natural language
- parametric model
- perplexity
- phoneme
- phonemes
- probabilities
- probability
- probability density
- probability density function
- probability distribution
- procedure
- process
- recognition accuracy
- recognition error rate
- recognition rate
- relation
- sentence
- sentences
- speech data
- spoken language
- statistical model
- test data
- test set
- tokens
- training
- training data
- training set
- tree
- triphone
- understanding
- word
- word classes
- word error rate
- word error rates
- word strings
- word-pair grammar
- words
