Pitt CRC contributes learning tools to cutting-edge teaching

High-performance computing resources of the Pitt Center for Research Computing enhance the productivity of Pitt researchers in many areas – in 2018, already as of May, the Center has worked with researchers in fields encompassing economics, political science and linguistics, in addition to researchers with an established history of using high-performance computing in fields such as chemistry, genomics and engineering.

One area of research unites all fields – improving teaching. Pitt CRC serves as a resource for faculty members to explore new methods to enrich student learning while preparing them for careers.

Na-Rae Han, PhD, Lecturer in the Department of Linguistics and Director of the Robert Henderson Language Media Center, steers her students to CRC computing clusters as part of her teaching practice.

“We wouldn’t describe what we do as cutting-edge research,” she explains. “This is cutting-edge teaching. We use the CRC computing cluster as a learning tool.”

Barry Moore ll, PhD, Research Assistant Professor at the CRC, worked extensively with Han to set up her students on the CRC clusters.

Han’s students do not need intensive computation power to work with data sets that typically don’t exceed 10 gigabytes. Instead, she believes her students need experience working in the command line computing environment to get a taste of Big Data.

“Linguistics students don't necessarily have a good sense of scale when it comes to data size,” she explains. “For those students, being able to crunch 10 gigabytes is a big deal. They are excited at the idea using supercomputers. And the CRC works to accommodate projects even at our scale.” She assigns student projects involving datamining in the Python language using a CRC cluster.

Experience in high performance computing enhances her students’ career prospects.

“Amazon and other companies working in natural language processing are hiring tons of computational linguists,” she explains. “Working on the cluster gives students familiarity with what they will see using linguistics data in their everyday jobs. As a linguistics teacher, it is wholly satisfying when the parents of one of my students ask, ‘What can you do with a degree in linguistics?’ – and then the student graduates and gets a great job. We’ve helped change their lives.”

Students from other departments come to the Language Media Center to learn to work with high- performance computing and big data. “We have become a gateway for other humanities students and faculty,” Han says. “They find us by word of mouth.”

Daniel Lambrecht, PhD, Assistant Professor in the Department of Chemistry, also employs CRC resources as a teaching tool. Lambrecht, who separately conducts his own research on CRC clusters, has been recognized many times, most recently by an American Chemical Society Junior Faculty Award in Computational Chemistry

For the fall 2017 term, Lambrecht developed a model of teaching a mathematics course for chemistry students with the support of a Pitt Discipline-Based Science Research Center’s (dB-SERC) Course Transformation Award. CRC provided Lambrecht with a service unit allocation and access to the Center’s Jupyter notebook server.

Lambrecht dubbed the project Mathematics on Computers for Chemistry Applications (MoCChA).

“This project addresses several needs. For one, computational modeling is an integral part of Chemistry research. In fact, the National Science Foundation has identified big data and computation as an education priority, and the CRC provides a perfect teaching platform. This class can function as a low threshold introduction to supercomputing and programming.”

“Chemistry majors often view physical chemistry as difficult because it relies on math. I think one of the challenges is to see how the abstract math formalism connects with the real world, and this is another need I want to address with this project. The Jupyter notebooks shift the focus from pencil and paper to doing mathematics by running computer simulations as ‘virtual experiments.’ This lowers the barrier for students to come up with ideas and to test them by running the simulations. Students are not expected to learn how to program, but they can modify simulation parameters and test their intuition by asking questions like ‘What would happen if we add heat?’  It ties the math to an actual Chemistry problem, and the cluster becomes a virtual lab.”

Lambrecht wants his students to enjoy multiple perspectives and multiple channels for learning.  Students work with the Jupyter notebook in teams of two or three, where students bring in their individual, often complementary skills. A percentage of grades are based on these team projects.

Student responses to the class were overwhelmingly positive, with an 85 percent approval rate among students answering a post-final survey. They described the Jupyter notebooks as helping them solve math problems while also focusing on the big picture. Building on his experience and student reactions, Lambrecht will teach the class using a refined model in the fall of 2018.

The students’ experiences, Lambrecht believes, may ultimately bring other researchers to use Pitt CRC resources. “If using the Jupyter server helps students through a challenging class, they will remember that the CRC is available. Some PIs might become connected to the Center through the awareness of their students.”

Moore at Pitt CRC also worked closely with Lambrecht. He is enthusiastic about the possibilities of incorporating high-performance computing into teaching and learning.

“Most educational uses of our clusters take up only a small fraction of the CRC’s computing power,” says Moore, who regularly conducts workshops and one-on-one teaching sessions. “Using our resources for education doesn’t take away from CRC’s high-performance computing research mission. In fact, teaching and learning enhance our mission. We play a role in helping grow the research community from the ground up.”

Contact:
Brian Connelly
Coordinator of Outreach and Communications
bgc14@pitt.edu