Finding a path to pursue your passion in research

PhD candidate Mohamed El Banani has always been interested in questions about perception and intelligence, which has led him to the field of computer vision and machine learning to build models that can understand images and video. 

We sat down with Mohamed to learn more about his experience as a graduate student at CSE. Here’s what he had to say:

Discovering computer vision 

I’ve always wanted to investigate how the mind works and how we perceive what we see everyday. I eventually decided that computer science was the best way to pursue those interests. 

Research fields like psychology and neuroscience have specifically focused on how humans perceive the world and how they understand it. However, those are very difficult questions to answer since we are limited in what experiments we can run. I think if we take a step back and consider how perception or intelligence can even work, then we can build computational models to directly emulate those capabilities and run experiments that we cannot perform with humans.

When I was working as an undergraduate researcher at Georgia Tech, one of the postdoc fellows suggested I consider Michigan for my graduate studies, and they recommended several faculty, one of which was Prof. John Laird, who was a pioneer in cognitive architecture research and was very interested in the intersection of cognition and computing. Once I arrived here and worked for John, I became more interested in how cognition works but also realized that my real passion was in computer vision.

Why U-M?

One of the main benefits of Michigan is the faculty’s diversity of interests. We have a lot of faculty who study very different topics. One challenge for me during grad school, for example, was that I wasn’t really sure what I wanted to focus on. I knew I was interested in several very general questions in artificial intelligence, but there are many ways that you can tackle that. You can focus on how cognition works, how perception works, how language works. Just having access to faculty that study those different problems at Michigan has been very good for me as I was able to talk to different people as I was figuring out my research focus. 

I’m advised by Prof. Justin Johnson, and in his lab we all work on computer vision because that’s our interest, but we each focus on very different problems. Some people work on 3D vision, others work on vision and language. We’ve had a lot of flexibility in exploring different problems, discussing them, and finding things that we’re interested in. I’ve actually recently transitioned from work on 3D vision to exploring problems at the intersection of vision and language. 

Collaboration at CSE

There’s a lot of interaction between students in different labs. My most recent work was a collaboration with Karan Desai who is another PhD student in my lab. There are also several papers from the department that result from collaborations between different research groups. Within my lab, we’ve been very supportive of each other, and Prof. Johnson has been a big help. He’s always happy to talk about what I can do next or  what my options are and how to best pursue them. That support, both in research and in life, has been very important to me.

A lot of my friends from CSE are also branching out to other departments. My favorite aspect of graduate study has been the ability to explore different research ideas and areas. The thing that remains constant though is that you spend a lot of time talking to other students. Chatting with people to get their perspectives helps to improve your work and expand your research interests.

Advice for new grad students

Figure out what you want to pursue academically, but also be flexible. It’s important to spend a good amount of time thinking about what you’re trying to get out of the program and what topics you are interested in.  By spending this extra time to figure out what you want, you might end up finding faculty in a school that you weren’t really considering, or figuring out that the topic that you thought you were passionate about is, in reality, very different from your expectations and that you’re actually interested in is something slightly different. 

By limiting yourself, you might end up applying to a program where there aren’t the resources to study what you want. Keep those things in mind and try to figure out the best program that will both address the subject that you’re really excited about, as well as give you flexibility to explore other topics. For example, a friend of mine began his research in computer science, but he figured out that he was actually more interested in computational linguistics. Fortunately, He was able to get a secondary advisor in linguistics and pursue interdisciplinary research without feeling limited to one area.