updates to process model exercises

[benpeloquin7]

• Some students found high variance between runs, which made for difficulty interpreting questions 1c-e. I think this typically had to do running the time function for too few samples. We can just set this to 5+.
• There were generally two modes here -- students who completely got stuff and those that were very lost / just didn't attempt parts of the problem. In particular, the idea of lifting the algorithms was difficult. Maybe we can provide better scaffolding...
• The more subjective 1F might be better worded as something like "what kind of inference algorithm do you think you use" and have the students introspect.