Code presentation improvements: Bayesian ML, Gaussian Distribution

[fonsp]

Working on:

• Gaussian Distribution (code changes, but not 100% happy with this one yet, because the example is not so spectacular. We should think of a new example together.)
• Probability Theory (no big changes)
• Bayesian ML 🌟 This one is completely restructured with more visible code, results, structure etc.

[bertdv]

Generally, I think it is a great update! Well done and thanks. Some specific issues for discussion on Bayesian ML Lecture (for each issue I added a screenshot of what I saw):

• Please move the optional slide on KL divergence to a mini. In general, important background materials that hinder the narrative of the lecture should preferably be moved to minis. In this case, teh KL divergence is important and I lke to refer to it from mulitple lecrtures, therefore I dont like it as an optional slide in the BML lecture.

• I prefer not to see any using package in the middle of the lecture notes. Can we move all using packages to the Appendix cell? Also, (optionally) perhaps use sliders in one of the graphs to change alpha and beta.

-[ ] I am happy with showing code per default for model specification, but I rather not interrupt the narrative to show other code, such as how many samples, and numerical values of the outcomes (BTW the outcomes should be 1's and 0's, not TRUE and FALSE, since we coded head as 1).

• Can we suppress Julia coming back with function definitions? EDIT: moved to "generic function with x methods" message #159

• I don't like to show numerical values for outcomes and inference results per default, especially for vectors. Perhaps we can make the below a "click for ..."-box instead? After inference, I like to show the results immediately by a graph.

• This is the graph we should show as results, immediately after inference:

• same here. Let's show the graph, but not the numerical evidences. Numerical values of evidence are also not important, it's about the comparison to evidence for another model.

**In short, I like what you have done very much, but I suggest the following for Code examples:

1. show code of model specification (and optionally refer to the equations in the lecture).
2. show code for inference procedure as optional: "Click for code of inference procedure"
3. show inference results in a graph (usually for the posterior). Ideally, with a slider to play.**

[fonsp]

I implemented the changes from our call, merging this 👍