How does the joint probability distribution help to generate things?
I am trying to understand the difference between discriminative models and generative models. One of the helpful answers at stack overflow is here: What is the difference between a generative and a discriminative algorithm?