ACL RD-TEC 1.0 Summarization of H89-2027
Paper Title:
THE N-BEST ALGORITHM: EFFICIENT PROCEDURE FOR FINDING TOP N SENTENCE HYPOTHESES
THE N-BEST ALGORITHM: EFFICIENT PROCEDURE FOR FINDING TOP N SENTENCE HYPOTHESES
Authors: Yen-Lu Chow and Richard Schwartz
Primarily assigned technology terms:
- acoustic recognition
- algorithm
- beam search
- beam search algorithm
- best-first search
- computing
- decoder
- discriminative training
- exhaustive search
- fast match
- hardware
- hidden markov
- hidden markov model
- markov model
- n-best paradigm
- n-best search
- natural language systems
- nl system
- processing
- pruning
- pruning strategy
- recognition
- recognition system
- scoring
- scoring answer
- search
- search algorithm
- speech recognition
- spoken language system
- spoken language systems
- time-synchronous viterbi-style beam search
- tuning
- viterbi
- viterbi decoder
- viterbi-style beam search
Other assigned terms:
- acoustic model
- acoustic models
- approach
- array
- beam
- case
- discourse
- distribution
- entropy
- error rate
- fact
- first-order model
- first-order statistical grammar
- frame
- generation
- grammar
- hypotheses
- knowledge
- language knowledge
- language model
- language models
- method
- n-gram
- natural language
- natural language knowledge
- nl model
- parse
- probabilities
- probability
- procedure
- process
- pruning threshold
- recognition component
- search problem
- search space
- search strategy
- sentence
- sentences
- spoken language
- statistical grammar
- statistical language model
- statistics
- syntactic model
- syntax
- syntax and semantics
- terms
- text
- theories
- theory
- training
- utterance
- vocabulary
- vocabulary size
- word
- word error rate
- word sequence
- word sequences
- words
