You are here
Rabih Younes: Integrating Machine Learning with Motion-Capturing Wearable Devices
September 12, 2018
New professor of the practice Rabih Younes blends the fields of wearable computing and machine learning

Rabih Younes is a new professor of the practice of electrical and computer engineering at Duke University specializing in using machine learning techniques to tease out detailed information about people’s activities from advanced wearable motion-capturing technologies. He earned his undergraduate and master’s degrees in computer engineering from Lebanese American University before completing his PhD in 2018 from Virginia Tech in the same field. Besides his research into data analysis from wearable devices, Younes brings a passion for teaching to his new position at Duke.
Welcome to Duke, Professor Younes! Can you tell us a bit about your research interests?
The work I did for my doctorate while at Virginia Tech was a blend between the two fields of wearable computing and machine learning. My advisor was designing a wearable system that could recognize your daily activities. The hardware was essentially a motion-capture suit, which entailed designing a loose-fitting sensor network and processing units that conformed to wearability constraints. I worked more on trying to blend the hardware with the software, which used machine learning and optimization techniques to classify the various activities of the user.
What’s the overall goal of developing such a system?
The motivation for the project is mainly in healthcare. When doctors need to track patients outside of the clinic, they rely on a log sheet where the patients are responsible for taking their physiological data and reporting what they were doing at the time. But if we could automate that entire process, not only would the data be much more accurate, it could be used to treat them better and predict what diseases they might develop at a later date.
Will you be continuing this line of research at Duke?
I won’t be working on that sort of a suit per se—I’ve always been more focused on using machine learning for the activity classification part of the process. In general, I’ll be working with motion capture tools toward developing systems that people would actually choose to wear on a daily basis and that would be much more comprehensive in their activity recognition than something like a Fitbit, for example.
As a professor of the practice, a large part of your work will be in teaching. What courses are you leading?
This fall I am teaching the ECE 350 course, called Digital Systems, which is the class I originally fell in love with that led me down this path of study. It has to do with the basic level of hardware design and sort of explains how you can build machines.
I’m also offering a new graduate-level course here at Duke that has to do with my research. It’s an ECE 590 course that is called Wearable and Ubiquitous Computing, which obviously has a lot to do with wearable technologies! Next semester I will continue teaching ECE 350 along with a graduate-level course—ECE 650—called Systems Programming and Engineering.
"I try to be as student-focused as I can. I try to motivate students in whichever ways work, even if I have to try a million different ways to motivate everybody and make them love the material."
What’s your philosophy when it comes to teaching?
I’ve actually always loved teaching—I’ve been doing it some form since I was a tutor in high school. And while I love my research, I actually got my doctorate degree so I could teach on the collegiate level. That’s why while I was at Virginia Tech I conducted some research in engineering education, which I plan to continue here.
Based on what I’ve learned from that work, I try to push my students to be better critical thinkers, and to communicate their science better with the world. These are two problems that are pretty much ubiquitous everywhere in the world. They manifest in different ways depending on where you are, but they are problems and they impact society a lot, so we should work on them.
And of course I try to be as student-focused as I can. I try to motivate students in whichever ways work, even if I have to try a million different ways to motivate everybody and make them love the material. After all, they can go to Google or register on something like Coursera and learn all kinds of things with these online tools, so I should give them a good reason to come to class.
Why did you choose to come to Duke?
Duke was one of the more highly ranked institutions that I applied to, and that was appealing. You can see the difference in Duke students—they’re the best of the best. It’s just my first week, but I think they’re amazing. They always want to learn more and have a complete understanding of every single detail, in the classroom and during my office hours. I expected something like that, but not to this level. But it’s good because it makes me come to class well-prepared to this challenging environment, and it’s great to have students that raise the bar this high. I bet that they’ll have some really nice projects at the end of the semester.