Clarify/standardize renewal model equations across docs

[cdc-mitzimorris]

This PR addresses all comments in issue #768 and further improves the discussion of the renewal models supported by PyRenew.

• It makes the temporal indexing explicit and distinguishes between the temporal indices for the renewal equation and the observation equation.
• It distinguishes between the mechanistic process and the stochastic model
• The discussion of latent temporal processes provides more nuanced explanation of the difference between the RW, AR1, and DifferencedAR1 temporal processes and the choice of hyperparameters for each.

[damonbayer]

Thanks @cdc-mitzimorris! I think we need to clarify "initial" vs "time 0". Other changes are more optional.

[cdc-mitzimorris]

@damonbayer I have made all requested changes:

• specify at time $0$ instead of "initial"
• refined plots for choice of temporal process adding 90% interval ribbon overlays with more discussion of what the prior predictive checks are showing
• replaced K as index on the generation interval with G, reserving K for subpopulation indices

[damonbayer]

Add a caption to this figure?

[damonbayer]

Use $\mathcal{R}(t)$ instead of Rt

[damonbayer]

You could actually say $\mathcal{R}(0)$

[damonbayer]

Thanks @cdc-mitzimorris! I requested a few more changes, but you can merge when you feel it is ready.