Discussion Forum

Hi Jacob,

So the notation that we used is based on the Bayesian networks. But it is not a class on graphical models, so you don't really need to understand the details of Bayesian networks and we almost exclusively only use linear chains in this course (i.e., the Markov Chains will not branch out).

I will advise you to read about its definition in the section about Data Processing Inequality on our class textbook (eq. 2.118-2.119 in my edition of Elements of Information Theory, Cover&Thomas)

By the way, if you want to, you can try to ask the professor on why he prefers to use an undirected notation during the exercise session. You will get an interesting discussion about it.

Hope it helps,
Best,
Reka