Lightly supervised training for risk-based discriminative language models.
Akio KobayashiTakahiro OkuYuya FujitaShoei SatoPublished in: INTERSPEECH (2013)
Keyphrases
- language model
- supervised training
- unsupervised learning
- language modeling
- probabilistic model
- conditional random fields
- training phase
- recognition process
- n gram
- information retrieval
- neural nets
- supervised learning
- query expansion
- dependency parsing
- training algorithm
- smoothing methods
- retrieval model
- semi supervised
- dependency parser
- machine learning
- higher order
- relevance model
- feature selection
- graphical models