On Thursday May 31st, 2018 we finally met with the research teams we will be working with. Part of the discussion was the scope of the project and what we, the REU participants (from here on out “interns”), want to do for the project over the coming 9 weeks. We immediately found an issue.
Leila and I want to work on the AI/Machine Learning part of TIMELI. Ahmed is uninterested in this line of work, and would more like to focus on the UI and HCI aspect of the TIMELI app.
What’s the big deal?
These two directions are orthogonal, as best I can tell. The ML runs in the background and provides a service that can (if implemented well on the frontend) assist the TIMs. The UI and HCI direction, among other things, includes the responsibility of tying the ML system to the TIMELI app frontend, but acts independently of backend implementation. This would not be an issue if we had the same freedom as a collection of three grad students. However we are interns in the same program and thus it is unlikely that the Dr. Gilbert and Dr. Winer will concede to splitting the team along lines of interest. This would go against the teamwork and problem resolution spirit of the program and would probably cause an undesired logistical headache for the next 9 weeks.
I should provide a deeper explanation of my motivation to pursue the ML direction. I intend to pursue a Master’s Degree in Computer Science with a focus on Intelligent Systems. I have done little work outside of the classroom that deals with applying ML and AI concepts, so I view this as a valuable opportunity to get hands on with real world machine learning. I am yet unfamiliar with doing hard work in an unstructured environment. I believe that this route would lead to working in a semi-structured semi-unstructured environment where I can learn how to work on open problems while still having resources (mentors) available to help and provide guidance. Additionally since the REU is semi-unstructured and I have not worked in depth with Machine Learning this route is a guaranteed challenge. And a good challenge is something I desperately desire.
I did come into TIMELI with the expectation it would be frontend work, and without the expectation that it would include backend work (ML). This is Ahmed’s experience as well. But where I am excited by the new opportunity, he is disinterested at best. I don’t view the UI and HCI direction with disinterest. Indeed, I would like to learn to design interfaces that facilitate the user in their accomplishing of their tasks. But as I emphasized earlier: the ML route is relevant to my long term interests, the UI/HCI direction is much less germane.