ACL RD-TEC 1.0 Summarization of P06-1078
Paper Title:
INCORPORATING SPEECH RECOGNITION CONFIDENCE INTO DISCRIMINATIVE NAMED ENTITY RECOGNITION OF SPEECH DATA
INCORPORATING SPEECH RECOGNITION CONFIDENCE INTO DISCRIMINATIVE NAMED ENTITY RECOGNITION OF SPEECH DATA
Authors: Katsuhito Sudoh and Hajime Tsukada and Hideki Isozaki
Primarily assigned technology terms:
- algorithm
- analyzer
- asr confidence scoring
- asr system
- automatic speech recognition
- character recognition
- chunking
- computational linguistics
- conditional random fields
- confidence scoring
- crfs
- cross validation
- cross-validation
- decision trees
- discriminant analysis
- discriminative named entity recognition
- discriminative training
- entity recognition
- error estimation
- feature selection
- forward-backward algorithm
- hidden markov
- hidden markov models
- hmms
- indexing
- information extraction
- information retrieval
- information retrieval and extraction
- japanese morphological analyzer
- kernel
- language processing
- linear discriminant
- linear discriminant analysis
- maximum entropy
- morphological analyzer
- named entity recognition
- natural language processing
- neural networks
- optical character recognition
- processing
- question answering
- recognition
- scoring
- search
- segmentation
- speech recognition
- spoken language processing
- support vector machines
- tokenization
- transcription
- validation
- viterbi
- viterbi search
- voting
- weighted voting
- word segmentation
Other assigned terms:
- acoustic model
- ambiguity
- annotator
- approach
- asr accuracy
- association for computational linguistics
- binary features
- broadcast news
- case
- character type
- chunk
- confidence measure
- confidence score
- confidence scores
- context words
- entropy
- estimation
- evaluation metrics
- f-measure
- feature
- feature value
- fmeasure
- generative models
- hypotheses
- hypothesis
- information sources
- language model
- language model score
- language processing tasks
- lattices
- linguistics
- markov models
- measure
- measures
- method
- named entities
- named entity
- names
- natural language
- natural language processing tasks
- ner model
- part-of-speech
- part-of-speech information
- part-of-speech tag
- perplexity
- posterior
- posterior probability
- precision
- probabilities
- probability
- procedure
- processing tasks
- prosodic features
- quadratic kernel
- segmentation ambiguity
- sentence
- sentences
- sigmoid function
- signal
- speech data
- speech signal
- spoken language
- support vector
- svms
- symbols
- technique
- terms
- test set
- text
- text corpus
- tokens
- training
- training data
- training set
- transcriptions
- trees
- vocabulary
- vocabulary size
- word
- word lattices
- word sequences
- words
