ACL RD-TEC 1.0 Summarization of W06-1314
Paper Title:
AUTOMATICALLY DETECTING ACTION ITEMS IN AUDIO MEETING RECORDINGS
AUTOMATICALLY DETECTING ACTION ITEMS IN AUDIO MEETING RECORDINGS
Authors: William Morgan and Pi-Chuan Chang and Surabhi Gupta and Jason M. Brenier
Primarily assigned technology terms:
- action item detection
- automatic generation
- automatic speech recognizer
- binary classification
- classification
- computational linguistics
- conditional random field
- crfs
- cross validation
- da classification
- error analysis
- feature analysis
- feature selection
- grouping
- hidden markov
- hidden markov models
- identification
- item detection
- learning
- learning techniques
- machine learning
- machine learning techniques
- markov model
- maxent
- maximum entropy
- maximum entropy model
- meeting recording
- named-entity recognition
- normalization
- optimization
- parameter optimization
- ranking
- reading
- recognition
- recognizer
- search
- segmentation
- speech recognition
- speech recognizer
- tagger
- tagging
- task management
- text classification
- utterance segmentation
- validation
Other assigned terms:
- 10-fold cross validation
- action item
- annotation
- annotator
- annotators
- approach
- asr output
- association for computational linguistics
- bigram
- bigram model
- binary classification task
- binary feature
- broadcast news
- case
- classification task
- classification tasks
- conditional markov model
- conditional probability
- contextual features
- conversational speech
- corpora
- detection task
- dialog
- dialog act
- discourse
- discourse markers
- distribution
- duration
- entropy
- entropy models
- f measure
- fact
- feature
- feature set
- feature sets
- feature weight
- feature weights
- frame
- generation
- human annotators
- hypothesis
- icsi meeting corpus
- implementation
- intention
- inter-annotator agreement
- kappa
- kappa statistic
- lexemes
- lexical features
- likelihood
- linguistics
- markov models
- maxent model
- maximum entropy models
- measure
- measures
- meeting corpus
- method
- methodology
- modal verb
- multi-party speech
- named-entity
- norm
- nouns
- oracle
- part of speech
- part-of-speech
- part-of-speech tags
- pauses
- pitch
- positive and negative examples
- precision
- probability
- prosodic behavior
- prosodic features
- punctuation
- semantic
- semantic features
- sequence model
- signal
- speaker change
- standard deviation
- statistic
- svms
- switchboard corpus
- symbol
- syntactic features
- syntactic patterns
- system performance
- tagging accuracy
- tags
- temporal expression
- temporal expressions
- term
- terms
- text
- timex
- training
- training data
- training set
- transcript
- transcripts
- unigram
- utterance
- verb
- word
- words
