ACL RD-TEC 1.0 Summarization of H92-1100
Paper Title:
SEGMENT-BASED ACOUSTIC MODELS WITH MULTI-LEVEL SEARCH ALGORITHMS FOR CONTINUOUS SPEECH RECOGNITION
SEGMENT-BASED ACOUSTIC MODELS WITH MULTI-LEVEL SEARCH ALGORITHMS FOR CONTINUOUS SPEECH RECOGNITION
Authors: Marl Ostendorf and J. Robin Rohlicek
Primarily assigned technology terms:
- acoustic modelling
- algorithm
- classification
- classifiers
- clustering
- context modeling
- continuous speech recognition
- feature analysis
- mean adaptation
- modeling
- modelling
- n-best rescoring
- phoneme recognition
- recognition
- rescoring
- score combination
- scoring
- search
- search algorithm
- segmentation
- signal process
- speaker adaptation
- speaker-independent recognition
- speech recognition
- split-and-merge search
- word recognition
Other assigned terms:
- acoustic models
- case
- continuous speech
- correlation
- duration
- error rate
- fact
- feature
- feature sets
- feature vectors
- formalism
- phoneme
- predictive power
- probabilistic framework
- probabilities
- process
- recognition task
- segments
- signal
- speaking rate
- stochastic model
- stochastic segment model
- test data
- test set
- timit corpus
- training
- training and test data
- weight space
- word
