General Probabilistic Networks:
  • An excellent review paper on learning probablistic networks from data by Wray Buntine, from IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING 1996.

  • Friedman, Murphy, and Russell show how to learn dynamic probabilistic networks from data in this paper.

Probabilistic Clustering:
  • Heller and Ghahramani show how to perform Bayesian Hierarchical Clustering which provides a probabilistic justification for clustering in a hierarchy, without requiring compute expensive techniques such as MCMC.

Bayesian Belief Networks:
  • Jeff Hawkins and Dileep George's excellent paper showing how they use Bayesian Belief Networks to perform classification, the first research results based on the theories expressed in On Intelligence by Hawkins.