ACL RD-TEC 1.0 Summarization of N06-1053
Paper Title:
TOWARDS SPOKEN-DOCUMENT RETRIEVAL FOR THE INTERNET: LATTICE INDEXING FOR LARGE-SCALE WEB-SEARCH ARCHITECTURES
TOWARDS SPOKEN-DOCUMENT RETRIEVAL FOR THE INTERNET: LATTICE INDEXING FOR LARGE-SCALE WEB-SEARCH ARCHITECTURES
Authors: Zheng-Yu Zhou and Peng Yu and Ciprian Chelba and Frank Seide
Primarily assigned technology terms:
- ad-hoc retrieval
- algorithm
- approximation
- audio\/video search
- automatic word alignment
- bottom-up clustering
- clustering
- computer science
- decoder
- decoding
- document ranking
- document retrieval
- grouping
- index representation
- indexer
- indexing
- information retrieval
- internet
- lattice search
- lvcsr
- matching
- navigation
- optimal pruning
- optimization
- phrase matching
- phrase spotting
- posterior-lattice indexing
- pruning
- pruning strategy
- ranking
- ranking method
- recognition
- recognizer
- relevance ranking
- retrieval engine
- retrieval system
- search
- search engine
- search engines
- searching
- segmentation
- software development
- speech indexing
- speech recognition
- speech recognizer
- spoken-document retrieval
- text representation
- text retrieval
- text search
- text-based information retrieval
- transcription
- user interface
- web search
- web search engine
- word alignment
- word spotting
- word-spotting
- word-spotting task
Other assigned terms:
- acoustic likelihood
- acyclic graph
- approach
- broadcast news
- case
- clustering procedure
- context free grammar
- continuous-speech
- corpora
- corpus size
- disk
- document
- document retrieval evaluation
- duration
- evaluation metric
- evaluations
- fact
- feature
- grammar
- hybrid word\/phoneme
- hybrid word\/phoneme lattice
- hypotheses
- hypothesis
- index
- inverted index
- keyword
- knowledge
- language-model context
- language-model context condition
- lattice
- lattices
- likelihood
- mean average precision
- meta-data
- metadata
- method
- nist
- phoneme
- phrase
- posterior
- posterior probability
- precision
- probabilities
- probability
- procedure
- process
- pruning threshold
- queries
- query
- query phrase
- query term
- recognition errors
- recursion
- relation
- retrieval task
- search results
- search time
- segments
- sentence
- sentence boundaries
- style
- system architecture
- technique
- technologies
- term
- terms
- text
- textual metadata
- theory
- transcript
- transcripts
- trigram
- user
- web audio content
- word
- word boundaries
- word error rates
- word hit
- word lattice
- word lattices
- word sequence
- words
- xml representation
