Zanyar:

• Don't mention specific technologies in the research subquestions
• Good to have some plan
• Boltzmann rationality

Javier Páez Franco:

• Maybe first questions are a bit too specific, may distract me from the main core of the project

Ivan

• Motivate well why my decisions.
• It is usually preferable to use already-used gym environments, as it would be easier to compare other algorithms with yours. If you use your own environments, explain why. Motivate your choices!
• It is completely fine if we don't use the library if it is too complex and overwhelming.
• In the final presentation, make sure you are not rushing. If you can't cover everything, you will need to choose.

Ana

• What challenges are you think about dealing with these subquestions? What metrics would tell you that your algorithm is doing good? Amount of queries, computational cost, etc.?
• Unpack the design of the algorithm, be a bit more specific with subtasks
• Some questions may be too hard or something like that to answer, we can combine them to answer them adequately in the paper

Petar

• How do you evaluate these things? Pearson correlation, etc., are possible options?

Midterm:

• May 22, 10 am tentative → We may change it if someone needs to do the resit of CAI (same day 9-12 am)