Add pipeline step for evaluation of all trained individual learners

[john-waczak]

We can output a quarto report with the graphs and associated table comparing evaluation metrics

[john-waczak]

• Include evaluation of predicted-true distribution closeness (e.g. KL-divergence)
• Hypothesis Tests (https://juliastats.org/HypothesisTests.jl/stable/)
• Mutual Information
• Correlation
• etc...

Basic question is how to define "statistical significance" for each our our modeling tasks.

[john-waczak]

https://arxiv.org/abs/1712.08048