ACL RD-TEC 1.0 Summarization of N03-1001
Paper Title:
EFFECTIVE UTTERANCE CLASSIFICATION WITH UNSUPERVISED PHONOTACTIC MODELS
EFFECTIVE UTTERANCE CLASSIFICATION WITH UNSUPERVISED PHONOTACTIC MODELS
Primarily assigned technology terms:
- 1-best recognition
- adaboost
- algorithm
- automatic speech recognition
- automatic training
- boostexter classifier
- boosting
- categorization
- classification
- classification algorithm
- classification method
- classifier
- classifiers
- database
- decision trees
- decoding
- finite state
- finite state automata
- hmms
- human-computer interaction
- interfaces
- iterative procedure
- labeler
- language modeling
- language modeling technique
- learning
- learning algorithm
- machine learning
- machine learning algorithm
- model training
- modeling
- modeling technique
- n-gram re-estimation
- natural language systems
- phone recognition
- phone string classification
- phone-string utterance classification
- processing
- re-estimation
- recogition
- recognition
- recognition system
- recognizer
- retrieval system
- retrieving
- speech processing
- speech recognition
- speech recognition system
- speech synthesizer
- spoken language system
- stochastic language modeling
- support vector machines
- synthesizer
- text categorization
- training procedure
- training process
- transcription
- unsupervised training
- utterance classification
- viterbi
- word insertion
- word recognition
- word-based classification
Other assigned terms:
- acoustic model
- acoustic models
- alphabet
- approach
- asr output
- automata
- baseline model
- classification accuracy
- classification model
- classification performance
- classifier model
- context information
- customer care
- data set
- data sets
- detection rate
- dialog
- discourse
- distribution
- duration
- error rate
- feature
- feature space
- language model
- language models
- lattice
- mapping
- method
- morphemes
- mutual information
- n-gram
- n-gram model
- n-gram models
- n-grams
- names
- natural language
- probability
- probability estimates
- procedure
- process
- recognition accuracy
- recognition model
- runtime
- speaking rate
- spoken language
- spoken natural language
- suffixes
- support vector
- switchboard corpus
- technique
- terms
- test data
- test set
- text
- training
- training data
- training examples
- training phase
- training set
- training time
- transcriptions
- trees
- trigram
- trigram model
- user
- user utterances
- utterance
- word
- word boundaries
- word level
- word string
- word strings
- word trigram
- word trigram model
- words
