Feature Markov Decision Processes (MDPs) [Hut09] are
well-suited for learning agents in general environments.
Nevertheless, unstructured ()MDPs are limited to rela-
tively simple environments. Structured MDPs like Dynamic
Bayesian Networks (DBNs) are used for large-scale real-
world problems. In this article I extend MDP to DBN.
The primary contribution is to derive a cost criterion that al-
lows to automatically extract the most relevant features from
the environment, leading to the "best" DBN representation.
I discuss all building blocks required for a complete general