Structure learning in hidden conditional random fields for grapheme-to-phoneme conversion.
Patrick LehnenAlexandre AllauzenThomas LavergneFrançois YvonStefan HahnHermann NeyPublished in: INTERSPEECH (2013)
Keyphrases
- structure learning
- hidden conditional random fields
- grapheme to phoneme conversion
- hidden markov models
- natural language processing
- bayesian networks
- graphical models
- visual speech recognition
- language resources
- conditional independence
- markov logic networks
- parameter estimation
- transfer learning
- network structure
- markov networks
- sample size
- machine learning
- probabilistic model
- probability distribution
- belief propagation
- maximum likelihood estimation
- machine translation
- random variables
- maximum likelihood
- worst case