ACL RD-TEC 1.0 Summarization of H05-1119
Paper Title:
SEARCHING THE AUDIO NOTEBOOK: KEYWORD SEARCH IN RECORDED CONVERSATION
SEARCHING THE AUDIO NOTEBOOK: KEYWORD SEARCH IN RECORDED CONVERSATION
Authors: Peng Yu and Kaijiang Chen and Lie Lu and Frank Seide
Primarily assigned technology terms:
- ad-hoc retrieval
- algorithm
- approximation
- audio recording
- browser
- computational linguistics
- database
- decoder
- decoding
- grouping
- human language
- human language technology
- index lookup
- indexing
- indexing approach
- indexing mechanism
- information retrieval
- inverted indexing
- keyword search
- keyword spotting
- language processing
- language technology
- large-vocabulary recognition
- lattice search
- linear search
- lvcsr
- mining
- natural language processing
- note-taking
- phonetic search
- processing
- ranking
- recognition
- recognizer
- relevance weighting
- retrieval system
- retrieval systems
- risk minimization
- search
- search process
- search system
- searching
- segmentation
- signal processing
- speaker segmentation
- speech recognition
- speech recognizer
- spelling
- spoken-document retrieval
- terminology
- text retrieval
- text-based information retrieval
- transcription
- two-stage search
- viterbi
- viterbi decoder
- weight pushing
- weighting
- word lattice search
- word-based representation
- word-lattice search
- word-spotting
- word-spotting task
Other assigned terms:
- acoustic likelihood
- acoustic model
- acoustic models
- acyclic graph
- approach
- association for computational linguistics
- backoff
- broadcast news
- case
- concept
- confidence scores
- conversation
- conversational speech
- corpora
- data set
- data sets
- dictionary
- dictionary entries
- disk
- distribution
- duration
- error rate
- evaluations
- experimental results
- formalism
- generation
- hybrid word\/phoneme
- hybrid word\/phoneme lattice
- hypotheses
- hypothesis
- implementation
- index
- indexing scheme
- key phrase
- keyword
- knowledge
- language model
- language model probability
- language models
- language-model context
- language-model context condition
- lattice
- lattices
- likelihood
- linguistics
- lvscr transcription
- measure
- measures
- method
- model probability
- names
- natural language
- nist
- oracle
- phoneme
- phoneme string
- phonemes
- phrase
- phrase boundary
- posterior
- posterior probability
- precision
- probabilities
- probability
- probability distribution
- procedure
- process
- pronunciation
- proper names
- queries
- query
- query phrase
- query term
- recognition errors
- recursion
- representations
- search results
- search time
- segments
- signal
- speaking style
- specialized terminology
- style
- switchboard corpus
- syllables
- symbol
- symbols
- system architecture
- technique
- technologies
- technology
- term
- term frequency
- terms
- test set
- text
- theory
- tokens
- topics
- training
- training data
- training set
- transcript
- transcriptions
- transcripts
- trigram
- trigram language model
- user
- vocabulary
- word
- word boundary
- word error rates
- word fragments
- word lattice
- word lattices
- words
