UNCG GraLNA REU: A Review

Tags: GraLNA REU, UNCG, Computer science, NSF REU, Graph theory, Machine learning

Introduction

At the University of North Carolina at Greensboro (UNCG), four professors in the Computer Science Department, led by Dr. Chunjiang Zhu (henceforth referred to as Dr. Zhu) and Dr. Qianqian Tong, applied-for and recieved a grant (#2349369) from the NSF to establish a Research Experience for Undergraduates (REU) program in 2024. As one of the members of the inagural cohort, I thought it would be useful to share my thoughts and experiences here, so that any future prospective students can have a better sense of the program, given that the program is expected to run until 2026.

Application Process

I know that REU programs can be competitive, and so my advice is to apply to as many programs as you can, especially as it does not cost anything to apply. Personally, I applied to around 50 REU programs, and recieved offers from 2 before accepting my offer from UNCG. However, unfortunately I was told to respond to my offer from UNCG by March 1st, which was over 2 weeks before the national (unoffcial) deadline of March 15th, which, as I understand it, is normally accepted by CS and Math programs. Luckily, by reaching out I was able to extend the deadline to March 3rd. Still, I didn’t have a chance to hear-back from a majority of the programs to-which I applied before I had to accept or decline my offer from UNCG.

To be honest, I was on the fence about accepting my offer until I got an email from the co-PI, Dr. Zhu, where he told me about how he liked my profile and was interested in working with me during the course of the program. I really appreciated this personal touch in the recruitment process, and it gave me a good feeling that the program would be a valuable experience. As it turns out, my intuition was quite correct.

Arriving in Greensboro

I booked a flight to the Piedmont Triad International Airport (GSO), which is a very small airport located just about 20 minutes from UNCG, arriving a day before the program began. The program graciously offered to cover $522 (I’m not sure how they picked this number) for roundtrip transportation to-and-from UNCG, which was enough if you were traveling domestically. However, be advised that you will only recieve reimbursement well after the program ends. I would expect not to be paid until around 6 months after the program ends, to be on the safe side.

After I arrived, Dr. Zhu picked me up from the airport and drove me to campus. He showed me to my dormitory in the Ragsdale/Mendenhall Building, which was nothing extraordinary, but on the upside, each student had their own, single room. As a word to the wise: bring your own bed sheets, don’t buy the temporary bedding on campus. The bed sheet package offered for about $10 to rent is quite horrible, and you surely want to be comfortable.

The campus itself is quite nice. Built in 1892, it was originally the all-women’s component of the University of North Carolina system, and some of that history is still visible on-campus. In the summer, the campus is very quiet and very empty. Don’t expect to be attending social events every day and going to new sites and venues. That said, I felt it was a good environment to focus on research, which is the main point of the program, after all.

Structure of the Program

The GraLNA REU is an 8-week program that is highly research-focused, with some group activities interlaced throughout. In summary, I would say that the program is very loosely structured, and you get out of it only what you put in, in-terms of effort. The first day, our entire cohort, consisting of approximately 10 students from universities around the U.S., met together in one of the academic buildings. We awkwardly introduced ourselves as we eagerly awaited an introduction of what was to come of our summers. Then, the four professors arrived and introduced themselves. After a short introduction to the program, we were instructed to fill out some paperwork, and complete a research training online course.

The first week was quite slow. We had a tour of the campus, and each day there was a short presentation, maybe about an hour, by Dr. Zhu, giving a high-level introduction to some of the topics in the field of graph theory and graph learning to probe our interest. The week culminated in each of the four professors giving an introduction of their research, and us selecting the professors with-whom we wanted to work. Not everyone got their first choice of advisor, but in the end I think everyone was relatively satisfied with their projects and team.

From then on out, each day was more or less the same (loosely structured, remember?). In the morning at 9am, I walked over to the Computer Science Department, sat down in a classroom, and worked on my research problem. Personally, I was working alone, but most other students were in groups of 2. About once or twice a week, I had a meeting with Dr. Zhu (my advisor) to discuss new ideas that I had and brainstorm new approaches and problems, which was a welcome interruption to my productive, but isolating individual work.

The main distruptions to this day-in, day-out research were our midterm and final presentations. It was fun to learn-about the other projects that my peers were doing, and to present about my work which I had been thinking very deeply about. I guess my main takeaway from the program is that this is an REU, and you will be doing research, every day. Daunting as that may be, especially if you’ve never done research before, looking back I feel there is no better way to learn than being thrown-in the deep end and struggling to swim to the surface.

Research Topics

Although I was focused on a topic that was very theoretical and mathematical, all of my cohorts were working on projects that were more applied. In hindsight, I think this program is really tailored to students interested in applied machine learning. If you are interested in theory, I would look-into Dr. Zhu’s research and see if it interests you. If not, then you’d better look elsewhere. My cohorts were working on projects in the areas of, for example, modeling the brain as a graph and applying deep learning tools, and new theoretical ML frameworks, namely, stochastic gradient descent. Personally, I was studying Differential Privacy (DP), which is a fun topic to look-into.

Amenities

First, let’s talk about the stipend. You are paid a handsome $700/week for your participation in this program, which is more than enough to cover living expenses. What’s more, you are given $600 for food, which you can spend anywhere (not just on campus), in addition to two group lunches per week. Although this is not enough to cover all of your food costs, it covers approximately half, eating out occasionally. Needless to say, I was able to have 1000s of dollars left over at the end of the program; a nice perk, indeed! The provided lunches are from the cafeteria, and are nothing to write home about, but who am I to complain about free lunch!

If you enjoy exercising, there is a brand-new, fanstastic gym that opened-up at UNCG, and you will have full, free access as part of the GraLNA program. Another nice perk is that there were several guest lectures incorporated into the program. Most of them were unextraordinary, expect for a presentation on how to give a quality, academic presentation, which was nothing short of life-changing. I felt like I was in a Ted Talk. The presenter, a Professor at UNCG, was quirky, funny, and one of the most excellent speakers I had ever had the pleasure of listening-to.

Our cohort was small, but we did manage to get along very well, thanks in-part to the group events hosted by Dr. Zhu. These included a talent show in our dorm, a trip to see a band off-campus, a campus tour of Duke University and UNC Chapel Hill, and a trip to the North Carolina Zoo in Asheboro. There was a close and intimate vibe between Dr. Zhu and us, since these events were more informal, and it felt like we were all just improvising and learning together as the program went-on. Sometimes, Dr. Qianqian Tong joined us for social events, as well.

Our cohort stuck together quite closely, especially as those without transportation were dependent on the students with cars to drive to the grocery store or to visit new places. Downtown Greensboro is within walking distance, about 1 mile away from campus, but is really not that interesting to visit more than a few times. I recommend visiting the Woolworth’s Building, in which the original bar which started the sit-in protests during the Civil Rights movement is preserved. Personally, I was really focused on my research question, and found that I spent most of my time thinking about it, even after the 5pm “clock-out” time. It was a bit lonely at times, but overall I felt sufficiently connected to the community, at least enough to last 8 weeks. This REU ran simulatenously with an REU from the Mathematics department at UNCG, so we were able to have some interaction with each other, though it was limited.

Conclusion

I will reiterate what I mentioned before: you will get out of this program only what you put in. Within our small cohort, I saw some students start the program very eager and excited to do an original research project, and end the program completely disconnected from their work and indifferent to the outcome. On the other hand, some students pushed through the difficult barrier to doing research in Computer Science and were very successful. In the end, I think the program had relatively high outcomes in 2024, with 4 papers being accepted to mid-level, but well-respected conferences, based solely on our work over the 8-week period. That is, a majority of participants ended the program with a published paper, many as lead authors. I have no doubt that future cohorts can achieve the same level of success. I would recommend the GraLNA REU to anyone who is interested in getting started with applied, computer science research, and wants to see if its something that they truly enjoy.

Please feel free to reach-out to me (you can find my contact on this website) if you have some more, specific questions about this program. However, I hope that this post has given you a pretty good idea about what to expect. Thanks for reading!