ACL RD-TEC 1.0 Summarization of H01-1059
Paper Title:
PORTABILITY ISSUES FOR SPEECH RECOGNITION TECHNOLOGIES
PORTABILITY ISSUES FOR SPEECH RECOGNITION TECHNOLOGIES
Authors: L. Lamel and F. Lefevre and J.-L. Gauvain and G. Adda
Primarily assigned technology terms:
- acoustic model adaptation
- acoustic model training
- acoustic modeling
- algorithm
- atis
- automated training
- bootstrap
- broadcast news transcription
- broadcast news transcription system
- classification
- continuous speech recognition
- conventional adaptation
- data base
- database
- decision tree
- decoding
- digit recognition
- direct recognition
- document retrieval
- em training
- estimator
- gender classification
- gender identification
- graph generation
- hmms
- human language
- human-machine dialog
- hypothesis generation
- identification
- information system
- internet
- iterative procedure
- language model training
- language modeling
- language transfer
- linguistic modeling
- map adaptation
- model adaptation
- model training
- modeling
- news transcription
- normalization
- partitioning
- recognition
- recognition system
- recognition systems
- recognition technology
- recognizer
- speaker adaptation
- speech recognition
- speech recognition systems
- speech recognition technology
- speech recognizer
- speech transcription
- spoken dialog
- spoken document retrieval
- spontaneous dictation
- spontaneous speech recognition
- supervised training
- text dictation
- training algorithm
- training procedure
- transcription
- transcription engine
- tuning
- unsupervised training
- vocabulary recognition
- word graph generation
- word recognition
- word recognizer
Other assigned terms:
- acoustic model
- acoustic models
- annotated corpus
- annotation
- annotators
- approach
- broadcast news
- broadcast news corpus
- broadcast news speech
- case
- cluster
- compound words
- compounds
- confidence measure
- continuous speech
- corpora
- data sets
- development set
- dialog
- discourse
- document
- duration
- error rate
- evaluation data
- evaluation set
- evaluation test
- evaluations
- events
- gaussian mixture
- generation
- grammar
- human annotators
- hypotheses
- hypothesis
- implementation
- interpolation
- interpolation coefficients
- key words
- knowledge
- labeled training data
- language model
- language models
- large corpus
- large text corpora
- lexica
- lexical coverage
- lexicon
- linguistic
- measure
- measures
- method
- n-gram
- names
- news corpus
- nist
- noise
- orthographic transcription
- perplexity
- phonemic transcription
- portability
- procedure
- process
- pronunciation
- recognition accuracy
- recognition corpus
- recognition task
- retrieval task
- seed
- segments
- sentences
- signal
- silence
- speaking style
- speech data
- speech recognition corpus
- spoken language
- statistics
- style
- supervised model
- system description
- system development
- system performance
- technique
- technologies
- technology
- television
- terms
- test data
- test set
- text
- text corpora
- topics
- training
- training data
- training material
- transcriptions
- transcripts
- transformation
- tree
- trigram
- trigram language model
- trigram model
- triphone
- typographical errors
- utterance
- vocabulary
- word
- word error rate
- word error rates
- word fragments
- word frequencies
- word order
- word sequence
- word sequences
- words