Dean's Medalist plans to pursue research in machine learning

Editor’s note: This story is part of a series of profiles of notable fall 2024 graduates.

Justin Le is an Arizona native, born in Glendale and raised in Peoria. While attending high school, he did not consider himself a “math person,” as he was more into writing and music. That changed when he took an advanced calculus class. The problems they were solving in calculus caught his interest. For the first time, he felt like he was creating something of substance when doing math, and this motivated him to study math in college.

This month, Le will be honored as the Dean’s Medalist from ASU’s School of Mathematical and Statistical Sciences (SoMSS). He will graduate summa cum laude from Barrett, the Honors College, with a bachelor’s degree in mathematics and a minor in applied computing. He will also receive the Moeur Award, given to undergraduates who maintain a 4.0 GPA.

Le has an outstanding 4.17 cumulative GPA and a 4.22 GPA in mathematics. As an honors student, he has excelled in some of the most rigorous courses, both in theoretical and applied mathematics. Notably, he has completed six 400-level depth courses — well beyond the two required for the degree. He is currently enrolled in four SoMSS courses, including a seventh depth course and three graduate-level core/qualifying courses, along with thesis hours for his honors thesis, "Diffusion Models for Generating Samples in Sparse Datasets," supervised by Associate Professor Sebastien Motsch.

“Justin has been conducting innovative research in machine learning, focusing on cutting-edge applications of neural networks. His first project demonstrated significant potential in biological imaging by developing segmentation algorithms that can automatically track and analyze the complex growth patterns of slime molds. This research has implications for better understanding biological systems. In his second project, Justin explored diffusion models, which are critical to modern generative AI technologies. By applying these models to fraud detection, he showcased the potential of machine learning to address real-world challenges in financial security. His work highlights the interdisciplinary nature of modern mathematics, bridging theoretical computational methods with practical problem-solving,” Motsch said.

Le plans to stay at ASU for one additional semester to finish his master’s degree through the 4+1 program. Looking ahead, he intends to pursue a PhD and continue his research in machine learning while also branching out into other areas such as partial differential equations.

“Justin is remarkably versatile and has a genuine enthusiasm for learning across disciplines. This curiosity makes him particularly well suited for fields like machine learning, where rapid advancements and experimentation drive progress. His adaptability and eagerness to tackle new challenges are qualities that will undoubtedly lead to meaningful contributions in any area he chooses to pursue,” Motsch said.

We asked Le to share his experiences as a Sun Devil.

Question: What is something you learned while at ASU — in the classroom or otherwise — that surprised you, that changed your perspective?

Answer: I found out rather quickly that research in math is not a clean endeavor, even though it may seem that way given how math research is presented. Many ideas get thrown around when doing research — sure, these ideas are rooted in some underlying theory, but at the end of the day, researchers are going by their gut feeling. As a result, the part of research that an audience will not see is the countless ideas that were attempted but did not work out as desired; groundbreaking results in math are built on the graves of many fruitless ideas.

Q: Which professor taught you the most important lesson while at ASU (and what was that lesson)?

A: I remember being in Intermediate Real Analysis I (MAT 472) taught by Dr. Brett Kotschwar, and he mentioned the importance of building one’s “short game.” In the context of that class, he meant that students should study by practicing many small problems to build basic skills in real analysis. However, I think that lesson strikes at something more fundamental: the importance of getting your reps in. Books, articles and videos are not education — they are just there to explain information. Education is attained by the individual, and it must come by repeatedly doing what you want to learn, whether that be academics, sports, art, etc.

Q: What do you like most about mathematics?

A: I’ve always enjoyed role-playing games. The feeling of exploring new places and slowly becoming stronger/more knowledgeable always appealed to me. Doing math gives me a similar feeling — when you start out studying math, you don’t know much and everything seems quite arcane. However, through lots of trial and error, you slowly clear the fog and acquire new tools that make you feel powerful. Treating math as exploration has made it so much fun for me, and the best part is that math is a game that cannot be completed — there is always something new to explore!

Q: What is the best piece of advice you would give to those still in school?

A: Get a whiteboard — the biggest whiteboard you can find — and hang it up on your wall. Now, when you’re studying, you get to role-play as a professor! Write down the things you want to study and explain them (out loud!) as if you were a professor giving a lesson. If you’re doing some practice problems, write out and verbally explain your steps, even if there is no one in the room with you. Of course, professors teach because they are brilliant, but I believe the converse is also true; professors are brilliant because they teach! If you can get to a point where you can explain ideas clearly, truthfully and in depth, you are much better equipped to succeed in your classes. To this end, having friends who can check your work is also crucial — they can even serve as your audience for your mini-role-play lessons!

Q: What was your favorite spot on campus, whether for studying, meeting friends or just thinking about life?

A: On the second floor of Noble Library, there’s a quiet room where people can come and study in silence. This place became a sanctuary for me to get away from the bustle and noise of Tempe. Whenever I needed to study on campus, I always found myself coming back to that room.

Q: What do you like to do in your spare time for fun?

A: For a long time, I’ve enjoyed cooking. Being able to make new dishes and try out new ideas has always been fun, and seeing the smiles of the people I feed is also quite fulfilling. Funnily enough, even though I’m supposed to be a math person, I don’t measure anything when I cook. I find it more fun to let my instincts and heart guide me (use your heart and not your eyes!).

Q: What do you think is most misunderstood about math by the general public?

A: That numbers are the most important part of math. When I say that numbers are not important to math, I’m not referring to how numbers get replaced by letters (i.e. when you learn algebra for the first time) — most people have already experienced that. Rather, I would say that math is best described as the study of patterns. In math, we observe patterns, make predictions about those patterns and then scrutinize our predictions. These actions are core to human thought, and every person does these things (whether intentional or not), even if they’re not thinking about numbers.

Q: If someone gave you $40 million to solve one problem on our planet, what would you tackle?

A: Educators are among the most valuable people on Earth; they teach, they inspire and they serve as role models for our communities. Because of this, I believe more should be done to financially support our educators. It is well known that our educators are not paid nearly enough, and action has been taken in the past to address this issue (see the Red for Ed strikes from 2018). Although $40 million is not sufficient to properly compensate every educator on the planet, starting small (i.e. in Arizona schools) may set a strong example and inspire change elsewhere.