Deriving conversation-based features from unlabeled speech for discriminative language modeling.
Damianos G. KarakosBrian RoarkIzhak ShafranKenji SagaeMaider LehrEmily Tucker Prud'hommeauxPuyang XuNathan GlennSanjeev KhudanpurMurat SaraclarDaniel M. BikelMark DredzeChris Callison-BurchYuan CaoKeith B. HallEva HaslerPhilipp KoehnAdam LopezMatt PostDarcey RileyPublished in: INTERSPEECH (2012)
Keyphrases
- language modeling
- discriminative models
- language model
- data sparseness
- information retrieval
- feature extraction
- query expansion
- unsupervised learning
- semi supervised learning
- finite state transducers
- probabilistic model
- feature vectors
- n gram
- classification accuracy
- retrieval model
- speech recognition
- class labels
- dirichlet prior
- statistical language models
- word error rate
- machine learning
- generative model
- text classification
- co occurrence
- feature space
- cross lingual
- natural language
- multimedia
- semi supervised
- information extraction