Professor Karan Ahuja is a man of many dimensions. He’s a Lisa Wissner-Slivka & Benjamin Slivka Assistant Professor in Computer Science and a member of Forbes 30 under 30. He also has a section on his personal website dedicated to “Fun Projects”–origami, glass-blowing and laser-cuts, to name a few.
Ahuja grew up in New Delhi, India before attending the Indian Institute of Information Technology in Guwahati. As an undergraduate, he interned at IBM Research and eventually attended Carnegie Mellon for his PhD in Human-Computer Interaction. He took a gap year at Google as a research scientist prior to arriving at Northwestern this fall. Now, he runs the Sensing, Perception, and Interactive Computing and Experience Lab, or the SPICE Lab for short, located on the third floor of Mudd Library.
It's a clean, organized space, with a couple of large machines in the corner for 3D printing and other manufacturing work. Next to an assembly of computer screens rests a plastic purple head. A box of purple t-shirts sits on a shelf above a wall of tools.
The lab specializes in technologies that “sense, track and understand humans to augment their interactions and assist them in their daily lives,” according to their website. This quarter, Ahuja taught a seminar on ubiquitous computing, a concept where computing technology is everywhere in a variety of formats, and hopes to teach a course on machine learning and sensing next.
“I took courses in networking that sounded fun. I did some database courses that sounded fun too. And I was really drawn towards application-centric research...so HCI really helped scratch that itch,” Ahuja said.
“We have so many smart devices around us, but these devices sort of know what’s happening around the world better than they know what you’re doing,” he explains. For example, a digital assistant like Siri is more likely to be able to inform you about the weather in Russia than your physical activity at that moment. Thus, the SPICE Lab focuses on building tools that help sense, track and understand humans. “We want to observe people in motion, rather than observing someone that’s in an MRI or a CT.”
When asked how his research interests have changed over time, he compares the graduate experience to being a blacksmith apprentice. “For example, before you make your own sword, or you make your own custom sort of design, you have to learn what is it that makes a good sword. So you learn your skills–you learn tempering, you learn working with metal,” he said. “Just like anything else, you have to learn the basics of what is good research.”
Ahuja notes that his interests have been molded by his advisors–initially at IBM and then Carnegie Mellon–before beginning to understand which elements of the work he found interesting and which ones he didn’t. Now that he’s a full-time faculty member, he’s found himself applying this in his own lab environment as well.
“You do your own research, but you also realize research through your students. And then there is a very interesting balance of making sure that students are on the right track and are doing the research that they care about, and marrying that passion with your research direction,” he said.
Taeyoung Yeon, a post-baccalaureate researcher in the SPICE Lab, started his year-long appointment at Northwestern after several years as a mobile game developer in South Korea. He's spending the year focused on research and now works on device sensing.
“The application is running itself on [a] smartwatch, so you can just detect what you’re doing right now using the audio and IMU sensors in this smartwatch,” Yeon said about his current project. His previous work experience sparked his interest in how mobile devices can affect people’s lives and provide meaningful feedback for health applications.
Ahuja co-advises Vasco Xu, a fourth-year doctoral candidate at the University of Chicago. Their collaboration began in the fall of 2023 when Xu was interested in the problem of indoor localization. Ahuja suggested using devices that people already have with them, such as phones and smartwatches, to help people locate themselves within an indoor environment.
“We’ve only put out a preliminary paper on that,” Xu said, “You can actually see yourself walking around in the environment. And so the immediate next step is, how can you localize yourself within the environment?”
Apart from pose and location tracking in indoor settings, the SPICE Lab is also currently interested in algorithms related to activity estimation. These research areas aim to provide insights to help people have better lives in terms of health outcomes.
“How can we make your workouts more productive? How can we make people who are in rehabilitation have better insights of their motion outside of clinic settings?” Ahuja said.
Working with biometric data comes with its concerns, and responsible management of the potential data procured from sensors is critical. Even from the conception of most of their technology, privacy is a priority. The lab tries to make its machine learning models run on the edge, or local devices and networks, rather than the cloud, which helps alleviate potential concerns of data mishandling by cloud providers and gives the researchers more agency and control over data integrity.
The researchers have also adopted using subsampled audio for some of their algorithms, recording at a low enough frequency so as not to capture the contents of a person’s speech. For example, they can capture audio at 1 kilohertz instead of 16 kilohertz.
“We can tell what you’re doing, maybe you’re chopping, or maybe you’re typing, or maybe you’re speaking,” he explains. “But we cannot tell what is inside that speech.”
Ahuja’s interest in tangible solutions isn’t just limited to computers. In his free time, he also enjoys working on hands-on artistic projects, exploring the Chicago food scene and playing squash.
“A lot of people say procrastination is bad. Creative exploration is bad. It’s a good way to remove yourself from the problem and meander a bit,” he said. “It lets you see the forest for the trees.”
Thumbnail by Yong-Yu Huang / North by Northwestern