Jordan M.I., Sejnowski. T.J. Graphical models: foundations of neural computation

MIT Press, 2001. — 433 p.Probabilistic Independence Networks for Hidden Markov Probability Models.
Learning and Relearning in Boltzmann Machines.
Learning in Boltzmann Trees.
Deterministic Boltzmann Learning Performs Steepest Descent in Weight-Space.
Attractor Dynamics in Feedforward Neural Networks.
Efficient Learning in Boltzmann Machines Using Linear Response Theory.
Asymmetric Parallel Boltzmann Machines Are Belief Networks.
Variational Learning in Nonlinear Gaussian Belief Networks.
Mixtures of Probabilistic Principal Component Analyzers.
Independent Factor Analysis.
Hierarchical Mixtures of Experts and the EM Algorithm.
Hidden Neural Networks.
Variational Learning for Switching State-Space Models.
Nonlinear Time-Series Prediction with Missing and Noisy Data.
Correctness of Local Probability Propagation in Graphical Models with Loops.