ACL RD-TEC 1.0 Summarization of N04-1017
Paper Title:
LATTICE-BASED SEARCH FOR SPOKEN UTTERANCE RETRIEVAL
LATTICE-BASED SEARCH FOR SPOKEN UTTERANCE RETRIEVAL
Authors: Murat Saraclar and Richard Sproat
Primarily assigned technology terms:
- algorithm
- asr system
- audio indexing
- automatic speech recognition
- cmu informedia
- collection-wide probability re-estimation
- continuous speech recognition
- decision tree
- document retrieval
- evaluation system
- factorization
- final state
- finite state
- finite state machines
- indexing
- information retrieval
- language model training
- large vocabulary continuous speech recognition
- length weighting
- lvcsr
- matching
- model training
- oov word detection
- phone recognition
- phonetic search
- phonetic speech retrieval
- phonetic transcription
- probability re-estimation
- pruning
- re-estimation
- recognition
- recognizer
- retrieving
- search
- searching
- speech recognition
- speech retrieval
- spoken document retrieval
- spotter
- switchboard evaluation system
- term weighting
- text retrieval
- text-to-speech
- transcription
- video mail
- weight pushing
- weighting
- word detection
- word recognition
- word spotting
Other assigned terms:
- acoustic models
- approach
- backoff
- broadcast news
- broadcast news corpus
- broadcast news type
- case
- composition
- confidence measure
- continuous speech
- conversation
- conversational speech
- corpora
- dictionaries
- dictionary
- document
- edit distance
- error rate
- evaluation metrics
- evaluation test
- f-measure
- fact
- feature
- hypotheses
- hypothesis
- index
- inverted index
- language model
- language models
- lattice
- lattices
- likelihood
- measure
- measures
- method
- mutual information
- news corpus
- nist
- particles
- phonemes
- phonetic representation
- phrase
- pointwise mutual information
- posterior
- precision
- probabilities
- probability
- procedure
- pronunciation
- pronunciation dictionary
- queries
- query
- recognition errors
- retrieval performance
- run-time
- segments
- speech corpus
- speech recognition errors
- switchboard corpus
- technique
- term
- terms
- test set
- text
- tokens
- topics
- training
- training corpus
- transcript
- transcriptions
- transcripts
- tree
- trigram
- user
- utterance
- vocabulary
- vocabulary size
- web site
- word
- word error rate
- word error rates
- word lattices
- word level
- word pair
- word strings
- word types
- words
