In the coming years, no technology is likely to impact as many career fields and areas of daily life as artificial intelligence (AI).
Learning how to work with it, then, is now a crucial part of a student’s educational experience – including in classes like assistant professor Srikanth Vemula’s Computer Science 317 course where 24 students have spent the past semester building, coding and training self-driving model cars using AI.
“We’re trying to take a hands-on approach – applying what students learn from how to design artificial intelligence agents, then taking that agent and applying it on the hardware,” Vemula said. “So we are combining the components of software development and hardware integration by introducing the concept of Sim2Real, which gives you an idea of how in the real world you might make the agent work from simulation to reality.”
The class met three times a week on Mondays, Wednesdays and Fridays. Students broke into teams using the open-source robotics platform DonkeyCar.
Each team was required to disassemble their DonkeyCar and replace certain parts with new ones using the 3D printers available in the Clemens Library Makerspace.
They then outfitted their car with a camera and coded programs that – through the use of AI – trained them to be self-piloting. The results of their labor were demonstrated on the final day of class (Dec. 9) in the Makerspace.
“In all of our classes – whether it be this one or our ethics course – everything we’re talking about is centered around a discussion of AI and how we can use and utilize it without having it overpower our own knowledge,” said CSB senior Sophia Maldonado, a computer science major. “Finding that balance is going to be the most important thing to know over the next few years, so getting the chance to work with the technology is really important.”
Those sentiments were echoed by SJU senior Joey Moberg, a fellow computer science major.
“Obviously, AI is something that’s going to continue to learn more and more on its own as time goes by, so learning how we can also implement new things will be extremely important,” he said. “For example, in this course, the path (the cars take) is mostly lefts and rights. But what if there’s something in the way? How does it recognize that and move around?
“That’s the kind of thing we’ve been learning.”
CSB junior Kendra Orbeck is a double major in both computer science and biochemistry on a pre-med track. She said the course offered perspective that will likely prove valuable as she pursues her career.
“AI is becoming more and more integrated into fields like health care, making it critical that we understand how it works,” said Orbeck, who this past summer worked with Maldonado on a Chatbot-centered research project.
“As a pre-health student, taking this course deepened my interest in exploring how AI can enhance patient care and gave me a better understanding of how its tools can be applied in the medical field.”
SJU senior Ben Hennen also saw broader applications for what he’s been learning.
“I worked in an AI department over the summer as an intern. So when I saw this class was being offered, it seemed like a way to continue developing those kinds of skills and to keep learning more,” said Hennen, also a computer science major. “I’m a data science minor too, so some of this runs very parallel to skills needed in data collection and data cleaning.
“It was a good mix of everything I like doing in one class.”
But Moberg said the process was certainly not all smiles.
“Honestly, the biggest lesson I’m taking away from this class is the need to be resilient,” he said. “We had a lot of problems with our hardware at first, and we couldn’t start on the software end until we got that figured out. When we finally did, we had to work hard because we only had three weeks left to get everything done.
“There were a lot of challenges, and it taught me the importance of hanging with it.”
That’s exactly the perspective Vemula was hoping students would come away with.
“I want them to be able to troubleshoot all aspects of this process, which will give them an overview of how the process of AI development works,” he said.
Vemula said a course like this would not be possible were it not for a lot of support from others in the campus community.
“I want to thank Adam Konczewski from Makerspace for the extensive collaboration he provided helping with the cars and spaces throughout the semester,” he said. “IT (Information Technology Services) for the network support, and my department chair Imad Rahal and (Associate Provost and Dean of the Faculty) Pam Bacon for encouraging me to do this.”
