Due: Thursday, Oct 17th 11:59PM
Grace: Sunday, Oct 20th 11:59PM
In class we talked about DuoLingo and we specifically posed this problem:
Challenge: The DuoLingo designers have come up with 50 different motivation interventions. Use an algorithm to decide when and for whom you can use the interventions. Note, a student could receive any number of interventions, even 0. You want to minimize how many students dropout from the platform and you care just as much about student dropouts now as in the future..
Specifically, once a day, for each student on duo lingo, you can chose to give any of the 50 interventions to the students. At any point in time you can see a students past experiences (their interactions on the website, which interventions they were given, and how many days since they were active). Each day you can observe which students have dropped out.
We don't expect that you already know how to approach this problem. Instead, we want you to see if you can discover an approach by doing a literature review (aka a google search).
(a) Spend 20 mins (or more) doing a literature review to find an algorithm to use to answer this question. Be creative. In your writeup include an explanation of what you searched for and where. In at least a few sentences describe your search strategy.
(b) After your literature review, what algorithm do you think would be useful for this problem? Write about 150 words, include at least one citation, and if possible include a diagram. It is fine if your algorithm is not "optimal". Try your best.
Brainstorm three ideas for your final project. Do this individually (eventually the final projects may be done in teams). Recall that final projects could either be research or a novel tool.
For each of the three ideas write an answer for the following questions:
Optionally do any work that you think would benefit yourself and/or the field of computational education. Report what you did. If you want to replace a problem on the pset with your work run it by us! Especially consider doing a short experiment that you could use as a proof of concept for any of the ideas your proposed in question 2