ACL RD-TEC 1.0 Summarization of N04-4022
Paper Title:
CONTEXT-BASED SPEECH RECOGNITION ERROR DETECTION AND CORRECTION
CONTEXT-BASED SPEECH RECOGNITION ERROR DETECTION AND CORRECTION
Authors: Arup Sarma and David D. Palmer
Primarily assigned technology terms:
- alignment technique
- asr system
- automatic speech recognition
- co-occurrence analysis
- context matching
- contextual analysis
- data analysis
- decoding
- dimensionality reduction
- error detection
- error detection and correction
- information retrieval
- information retrieval system
- language modeling
- language modeling approach
- matching
- matching technique
- maximum likelihood
- modeling
- phonetic alignment
- phonetic analysis
- post-processing
- processing
- pruning
- question answering
- recognition
- recognition systems
- retrieval system
- retrieving
- search
- searching
- speech recognition
- speech recognition systems
- speech retrieval
- speech system
- summarization
- unsupervised approach
- word co-occurrence analysis
Other assigned terms:
- approach
- asr output
- broadcast news
- broadcast news speech
- case
- co-occurrence
- co-occurrence information
- co-occurrence statistics
- co-occurrences
- conditional probability
- context window
- context words
- contextual features
- contextual model
- contextual similarity
- dimensionality
- document
- document collection
- error rate
- evaluation data
- fact
- information flow
- joint probability
- kullback-leibler divergence
- language data
- large corpus
- lexical co-occurrence
- lexical context
- lexical similarity
- lexicon
- likelihood
- measure
- method
- model parameters
- morphological variant
- mutual information
- n-grams
- natural language
- phonetic similarity
- pointwise mutual information
- precision
- probabilities
- probability
- process
- query
- query term
- recognition errors
- retrieval performance
- retrieval task
- search corpus
- similarity between words
- speech recognition error
- speech recognition errors
- spoken language
- statistics
- target word
- technique
- television
- term
- test corpus
- test data
- test set
- training
- training corpus
- training data
- training set
- transcript
- transcripts
- trigram
- user
- user query
- vocabulary
- vocabulary size
- window size
- word
- word co-occurrence
- word error rate
- word sequence
- words
