Probabilistic Graphical Model:
Representation:
Directed Acyclic Graph: Bayesian Network, HMM...
Undirected Graphical Model: Markov Random Field Ising Model, Conditional Random Field, Boltzmann Machine, Restricted Boltzmann Machine
Inference:
Exact Inference: Variable elimination, Junction-tree, Message-passing(belief propagation)
Approximate Inference: Loopy belief propagation Variational inference optimization-based, Monte-Carlo sampling-based
Learning:
Learn the Model Parameters, Graph Structure.
Maximum likelihood Estimation
Bayesian Inference for model parameters
Expectation Maximization Algorithm for latent variable / missing data