Model dependencies explicitly, then perform inference over the structure.

These pages break structured probabilistic models into individual lessons: factorization, conditional independence, message passing, visible Markov dynamics, hidden states, decoding, and parameter learning.