Mismatch between the initial discussion and the toy example in the second tutorial


#1
  • What tutorial are you running?
    intro_part_ii
  • What version of Pyro are you using?
    0.1.2
  • Please link or paste relevant code, and steps to reproduce.
    I just have a question about the math

The second tutorial begins by discussion latent variables and how they apply to bayesian inference.

However the toy example discusses a problem involving a noisy scale that can be modeled using a univariate gaussian. However, the model discussed has no latent variables. It does, however, have a prior that is also modeled with a univariate gaussian and that is fine.

I feel that the initial discussion is a bit misleading and doesn’t really lend itself to the rest of the tutorial. Why not discuss priors, when they are being used by the model?