Approximate inference: A sampling based modeling technique to capture complex dependencies in a language model.
Anoop DeorasTomás MikolovStefan KombrinkKenneth ChurchPublished in: Speech Commun. (2013)
Keyphrases
- language model
- approximate inference
- dynamic bayesian networks
- graphical models
- probabilistic model
- language modeling
- belief propagation
- n gram
- probabilistic inference
- exact inference
- parameter estimation
- markov chain monte carlo
- bayesian networks
- latent variables
- retrieval model
- gaussian process
- information retrieval
- message passing
- conditional random fields
- dependency structure
- mixture model
- query expansion
- smoothing methods
- structured prediction
- gibbs sampling
- markov networks
- belief networks