Slides 20.1. Statistical Learning 20.2. Learning with Complete Data Maximum-likelihood parameter learning: Discrete models Naive Bayes models Maximum-likelihood parameter learning: Continuous models Bayesian parameter learning Learning Bayes net structures 20.3. Learning with Hidden Variables: The EM Algorithm Unsupervised clustering: Learning mixtures of Gaussians (Discrete hidden variable, Continuous evidence) Learning Bayesian networks with hidden variables () Learning hidden Markov models (Learning over time) The general form of the EM algorithm Learning Bayes net structures with hidden variables 20.4. Instance-Based Learning Nearest-neighbor models Kernel models 20.5. Neural Networks Units in neural networks Network structures Single layer feed-forward neural networks (perceptrons) Multilayer feed-forward neural networks Learning neural network structures 20.6. Kernel Machines 20.7. Case Study: Handwritten Digit Recognition |