Reconciling expert opinions example

[BenZickel]

This tutorial demonstrates how time dependent expert opinions can be reconciled in a Bayesian framework, by using the following effect handlers in order to modify model behavior:

• poutine.reparam for splitting observations to several random variables using the SplitReparam reparameterizer.
• poutine.condition for assigning fixed values to random variables.
• poutine.equalize for forcing random variables to have the same values.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[martinjankowiak]

thanks! can you please add CI logic using smoke_test as demo'd e.g. in this notebook? this way the nb will get tested

https://github.com/pyro-ppl/pyro/blob/dev/tutorial/source/logistic-growth.ipynb

[BenZickel]

Thanks for the review @martinjankowiak! I've added the smoke test logic.

[martinjankowiak]

looks great!

• typo: conditined
• explain args to SplitReparam?