Dynamic Learning with the EM Algorithm for Neural Networks.
João F. G. de FreitasMahesan NiranjanAndrew H. GeePublished in: J. VLSI Signal Process. (2000)
Keyphrases
- em algorithm
- neural network
- hidden variables
- maximum likelihood
- expectation maximization
- learning algorithm
- mixture model
- parameter learning
- maximum likelihood estimation
- gaussian mixture model
- parameter estimation
- incomplete data
- generative model
- active learning
- multi modal
- supervised learning
- gaussian mixture
- prior knowledge
- training data
- log likelihood
- mixture distribution
- penalized likelihood