ACL RD-TEC 1.0 Summarization of W06-1644
Paper Title:
STYLE & TOPIC LANGUAGE MODEL ADAPTATION USING HMM-LDA
STYLE & TOPIC LANGUAGE MODEL ADAPTATION USING HMM-LDA
Authors: Bo-June (Paul) Hsu and James Glass
Primarily assigned technology terms:
- adaptive language modeling
- algorithm
- approximation
- automatic speech recognizer
- classification
- computational linguistics
- computer science
- computing
- context-dependent interpolation
- decomposition
- document modeling
- factorization
- gibbs sampler
- gibbs sampling
- hidden markov
- hidden markov model
- indexing
- language model adaptation
- language modeling
- language processing
- latent dirichlet allocation
- latent semantic analysis
- learning
- lecture processing
- lecture segmentation
- linear interpolation
- machine learning
- markov model
- matrix factorization
- maximum entropy
- mixture weight adaptation
- model adaptation
- model interpolation
- model selection
- modeling
- n-best rescoring
- natural language processing
- principal component analysis
- probabilistic lsa
- processing
- random sampling
- random selection
- recognition
- recognizer
- rescoring
- sampling
- segmentation
- semantic analysis
- singular value decomposition
- smoothing
- speech recognizer
- taggers
- text classification
- topic adaptation
- topic segmentation
- transcription
- vocabulary selection
- weighting
- witten-bell smoothing
Other assigned terms:
- approach
- asr transcript
- association for computational linguistics
- bias
- cache
- case
- cluster
- clusters
- collocation
- concept
- content words
- convergence
- conversational speech
- data structure
- development set
- dirichlet allocation
- dirichlet distribution
- distribution
- document
- domain model
- dynamic model
- entropy
- error rate
- feature
- feature vectors
- heuristic
- hypotheses
- implementation
- interpolation
- keyword
- knowledge
- labeled training data
- labeling
- language model
- language models
- large training
- latent semantic
- likelihood
- linear algebra
- linguistics
- local context
- maps
- markov chain
- matlab
- measure
- method
- model combination
- model parameters
- n-gram
- n-gram language model
- n-gram model
- n-grams
- natural language
- opinions
- paragraph
- parameter values
- part-of-speech
- perplexity
- perplexity reduction
- posterior
- prepositions
- priori
- probabilities
- probability
- probability estimates
- process
- recognition errors
- semantic
- sentence
- sentences
- signal
- statistics
- style
- style model
- style trigram model
- syntactic behavior
- syntax
- target sentence
- test set
- text
- textbook
- toolkit
- topic language model
- topics
- training
- training corpus
- training data
- training document
- training set
- transcript
- transcriptions
- transcripts
- trigram
- trigram model
- unigram
- unigram topic
- unlabeled corpus
- utterance
- vocabulary
- vocabulary size
- word
- word collocation
- word distribution
- word error rate
- word error rates
- words
