Course on knowledge graphs link students from across majors
In the summer of 2020, Distinguished Professor Isabel Cruz took a timely course about online teaching. Cruz, like fellow academics around the world, had been abruptly forced to shift her classes online due to COVID-19 and thought anything that could help her enhance the online learning experience for her students would be valuable.
“Part of our class focused on choosing a topic—it could be something you had taught before or a new course—and completely reimagining it online, from start to finish. You would recreate the whole thing,” Cruz said.
From that experience, the interdisciplinary CS586 Data and Web Semantics: “Knowledge Graphs” was born, and Cruz couldn’t have been more pleased with the turnout, attracting what she described as the perfect mix of master’s and PhD students from across the College of Engineering and beyond.
The class focused on data modeling and semantics, knowledge representation, querying, metadata, data integration and interoperation, web services, and applications.
Cruz split the class into five groups, each headed by a domain expert who was a PhD student from outside of computer science. They included three mechanical and industrial engineering students, a chemistry major, and a biomedical and health informatics student. Each domain expert was joined by at least two computer science students, typically one PhD student and one master’s student.
The teams worked on a project grounded in the domain expert’s research, helping them model their work.
Jackie Rojas Robles, a mechanical engineering PhD student who was the domain expert for her team, was intrigued when she heard about the course, and hoped to build a tool similar to one she uses on Nanohub that can more easily locate data on antimicrobial coatings for metals, which will let her simulate other properties of metals.
“With help from my team members, I wanted to explore how to structure that knowledge, maybe facilitate the design and use of these coatings more easily,” Rojas Robles said.
Lydia Tse, a computer science master’s student and data visualization engineer for Nike, was attracted to the interdisciplinary aspect of the course.
“It seems like we seldom have the opportunity to interact with other fields and take the skills we’ve learned to solve other problems,” Tse said. “It was valuable on multiple levels.”
Tse worked on a project related to COVID-19 data, using natural language processing techniques to distill information.
“Working with someone who was not perhaps as well-versed in computer science techniques, I had to bridge that gap and present the information in a way that is coherent and would let them ask questions,” Tse said. “It was good practice; how do you get to know your audience when presenting highly technical information, and how do you distill it so they understand enough so you can move forward together?”
Cruz agreed, noting the challenges of working across disciplines, which she feels is at the heart of data science.
“The most difficult problems I have to solve are from other disciplines,” Cruz said. “As a researcher, maybe I’m working with someone in geography, or the College of Medicine. They know what they want to do but not how to do it. And I have to help figure out how to help them represent the data and work with their data.”
Cruz admired the hard work of her students, praising their dedication and curiosity, saying they took to the material fearlessly.
“What I enjoyed the most was that it was structured to let everyone deep dive into the topic they were interested in and collaborate with people you don’t usually work with,” said Rojas Robles.