By Aida Rahmattalabi, Caroline Johnston, and Phebe Vayanos
On college campuses, racial minorities receive 22% of bachelor and 9% of doctorate degrees [1]. In the workforce, those who identify as Black and/or Hispanic account for 16% of the STEM field, while making up 27% of the U.S. population [2]. With funding from the INFORMS DEI Ambassadors program, our team came together wanting to expose high school students from underrepresented backgrounds to the fields of operations research (OR) and artificial intelligence (AI). We, Dr. Phebe Vayanos and PhD students Aida Rahmattlabi and Caroline Johnston, are members of the University of Southern California (USC) Center for Artificial Intelligence in Society (CAIS), whose primary goal is to conduct research in AI/OR to tackle some of the most difficult societal problems of our time, especially those affecting underserved populations and underrepresented groups. For this reason, we wanted to create an event in which students not only learned about technical aspects of these fields, but also how they can be applied to problems of societal importance - problems these students see facing our world today. Even further than this, these students should be able to connect to and build a network with researchers who are active in this domain.
Piggybacking on the INFORMS Annual Meeting, we planned to host the event on the Saturday before the start of the meeting. Students would work on small team projects with world-class mentors from the OR/AI fields, culminating in group presentations. Given that INFORMS had originally been set to take place in National Harbor MD, we started reaching out to different organizations and schools in the area. Utilizing the resources at USC, we contacted Katie Mills, the co-director of the USC K-12 STEM Center who was able to connect us with an organization to partner with near National Harbor, MD - Code in the Schools. Code in the Schools is an organization with the mission to expand “access to quality computer science education” to youth in the Baltimore area and specifically those from traditionally underrepresented groups in technology fields. This was perfectly aligned with our proposed effort!
The planning process happened in early March. The world then turned on its head due to the pandemic and we suddenly needed to reimagine this event in a virtual format. Over the summer, we ourselves attended various online conferences, experimenting with which formats seemed to work best. How could we virtually engage students (on a Saturday!) while still carrying out our original mission? We eventually found Gather.town, an online video game-esque environment where users control a virtual avatar to interact with one another - a seemingly great medium for engaging the high schoolers of today.
Though virtual interactions are not always preferred to those that are in-person, a few silver linings came with hosting the event online. No longer limited by a physical location, we were also able to involve students from the Los Angeles-based STEM Academy of Hollywood to create a cross-coast event. Similar to the Code in Schools, a majority of the students at the STEM Academy of Hollywood come from racial groups that are typically underrepresented in STEM which created a great opportunity to reach more students than we had originally planned. In addition to our student pool, now faculty and Ph.D. students from across the continent from USC to Georgia Tech to the University of Toronto were able to participate as mentors as well.
On Saturday, November 7th, we hosted our 3-hour event entitled, “explOR”, in which small groups of ~3-4 students (mixed between Baltimore-based and Los Angeles-based) were assigned a mentor in the OR/AI fields. The mentors presented their group with a problem of societal importance ranging from mitigating human-wildlife conflict, optimizing public defibrillator placement, or increasing the access of homeless youth to housing resources. The students then engaged in a virtual “scavenger hunt” in which they explored the Gather.town environment and interacted with the other mentors. Each mentor used slides or a video for a high level explanation of an OR/AI technique and its applications. These methods ranged from game theory and linear programming to machine learning and computer vision.
Then students regrouped with their mentor and devised a plan for how to apply these learned techniques to their project. Students created a presentation of their plans and presented to the entire group at the end of the day. The results were very impressive! Utilizing the Zoom “poll” feature, all participants voted for the “Best Presentation,” for which the winners received Amazon gift cards (though in the end, all participants received a gift card for their enthusiasm and engagement)!
From our post-event survey, ~41% of students thought that the most enjoyable part of the event was getting to work with a mentor who currently does research in OR/AI on problems of societal importance. After the event, 100% of students said that they are “more interested in one or more specific fields of engineering and/or computer science, such as OR or AI.” In total we received 17 survey responses out of 22 participants, which results in a response rate of ~77%.
You can find our website here to learn more about the wonderful mentors who were instrumental in the event’s success and the projects they guided the students towards solving.
We would additionally like to thank Elizabeth Bondi for her helpful conversations in planning this event through her previous experience in running the “tryAI” initiative.
[1] National Science Foundation (NSF), National Center for Science and Engineering Statistics 2019
[2] Pew Research Center report on Diversity in the STEM 2018
Aida Rahmattalabi, 5th year Computer Science Ph.D. student at the USC
Caroline Johnston, 2nd year Industrial & Systems Engineering Ph.D. student at USC
Phebe Vayanos, Assistant Professor of Industrial & Systems Engineering and Computer Science and Associate Director of USC’s Center for Artificial Intelligence in Society (CAIS)