Showcase Features Undergraduate Student Summer Research Projects

Research experience is an integral part of advanced undergraduate education, especially for students planning to enter graduate school.

Over the summer, a cohort of Northwestern Computer Science students conducted practical, immersive research in labs alongside graduate students and faculty members.

On October 17, five of the summer researchers presented their contributions to new knowledge during the Undergraduate Research Showcase.

Sara Owsley Sood, Chookaszian Family Teaching Professor of Instruction and associate chair for undergraduate education at Northwestern Engineering, co-organized the event with student affairs coordinator Bella Barrios.

The Department of Computer Science supported 24 student research projects this summer. Three students also received support from the US National Science Foundation Research Experiences for Undergraduates program through funding received by Peter Dinda, professor of computer science and (by courtesy) electrical and computer engineering at Northwestern Engineering.

Samir Khuller, Peter and Adrienne Barris Chair of Computer Science at the McCormick School of Engineering, provided the welcome remarks and discussed the impact of research.

"We offer a wealth of research opportunities to our students here at Northwestern CS,” Khuller said. “It's really a joy to see so much interest in our new research track for undergraduates. These opportunities allow students to delve deep into an area, well beyond where their normal course work might take them."

Defne Deda
Summer research adviser: Mike Horn

Deda is a fourth-year student pursuing a degree in computer science and a minor in psychology through Northwestern’s Weinberg College of Arts and Sciences. She is also earning a Segal Design Certificate at Northwestern Engineering’s Segal Design Institute.

Deda presented "Stamp-tastic," an interactive web tool designed to introduce third- to fifth-grade students to key computational ideas by creating and modifying digital art stamps.

“By combining art with foundational coding principles, the tool fosters creativity, problem-solving, and logical thinking in a familiar and accessible environment,” Deda said. “Early results from our pilot tests demonstrated the effectiveness of this visual and interactive method in demystifying computational concepts, providing a scalable prototype for future use in diverse educational settings.”

Didier Munezero
Summer research adviser: Christos Dimoulas

Munezero, a second-year student in computer science, explored the feasibility of using large language models (LLMs) to automate unit testing in software development. His team investigated whether LLMs can be utilized to both simplify testing and build effective testing platforms.

Munezero and his team developed a testing tool for beginner-level Racket programs that synthesizes a series of prompts for GPT 3.5, increasing the complexity of the context and information with each prompting round.

Nathan Pruyne
Summer research adviser: Bryan Pardo

A member of Bryan Pardo’s Interactive Audio Lab, Pruyne is studying computer science at Northwestern Engineering and music technology in Northwestern’s Bienen School of Music.

Pruyne examined attribution for generative music models. He presented a system easily integrated into a musician’s workflow that detects similar segments of music based on both melodic and timbral aspects.

“With the rapid growth of generative music models, attributing the outputs of these models to parts of their training data has become increasingly important both for presenting users with more insight on their usage and for avoiding copyright infringement,” Pruyne said.

Pardo is a professor of computer science at Northwestern Engineering and a codirector of the Center for Human-Computer Interaction.

Marko Veljanovski
Summer research adviser: Zach Wood-Doughty

Veljanovski is a fourth-year student pursuing a double major in computer science and mathematics. He conducted research this summer on invariant risk minimization (IRM), an approach for out-of- distribution generalization using an optimal classifier that remains consistent across environments.

“While IRM has been extensively tested with image data, text-based datasets remain underexplored, despite out-of-distribution generalization being crucial for large language model performance,” Veljanovski said.

Veljanovski and his team designed and adapted a flexible synthetic text data generating process to evaluate IRM for natural language processing.

Benjamin Ye
Summer research adviser: Simone Campanoni

Ye is a third-year student in computer science at Northwestern Engineering. An aspiring software engineer, Ye is a member of Simone Campanoni’s Rethinking Compiler Abstractions for New Applications (ARCANA) Lab.

Ye is developing a Rust language extension and compiler front-end for a Memory Object Intermediate Representation (MemOIR) system. The Rust-MemOIR system uses properties of single-ownership memory to generate novel optimizations for languages that manage memory manually.

“Rust's borrow checker moves the burden of alias analysis from the developer to the compiler and allows us to more paradigmatically utilize the range of optimizations offered by MemOIR,” Ye said.

Campanoni is an associate professor of computer science and (by courtesy) electrical and computer engineering at Northwestern Engineering.