This is not naive bayes and does not assume independent observations on the exercises. The point of using a network is to model the joint distribution with dep

This is not naive bayes and does not assume independent observations on the exercises. The point of using a network is to model the joint distribution with dependencies.