Short answers, pulled from the story.
Judea Pearl coined the term Bayesian network in 1985. He chose the name to emphasize the subjective nature of input information, reliance on Bayes' conditioning for updating beliefs, and the distinction between causal and evidential reasoning.
A causal network is a special case of a Bayesian network that requires each directed edge to represent a genuine causal relationship. In a standard Bayesian network, an arrow from one node to another indicates a conditional dependency, not necessarily a causal one.
No. Cooper proved in 1990 that exact inference in Bayesian networks is NP-hard. Roth further showed it is #P-complete, as hard as counting satisfying assignments of a logical formula, and approximate inference is also NP-hard except under specific restrictions.
Pearl's Probabilistic Reasoning in Intelligent Systems and Neapolitan's Probabilistic Reasoning in Expert Systems, both from the late 1980s, summarized the properties of Bayesian networks and established them as a formal field of study.
A dynamic Bayesian network models sequences of variables, such as speech signals or protein sequences. It extends the standard Bayesian network framework to represent dependencies that unfold over time.
The Markov blanket of a node consists of its parents, its children, and any other parents of its children. This set renders the node conditionally independent of all other nodes in the network, so knowledge of the Markov blanket is sufficient to determine the node's distribution.