Semester Award for Tarek Alakmeh
Tarek Alakmeh has been awarded a Semester Award for his outstanding Master’s Thesis “Decoding a Developer’s Mind: Multimodal Approaches to Code Comprehension” (Fall 2024). Congratulations!
Modern society relies on software developers to maintain the systems we use every day. However, efficiently comprehending and evolving code remains a central challenge. With the growing use of generative AI, reviewing and comprehending code may matter even more than writing it. Traditional ways of reasoning about comprehensibility, such as decades-old complexity metrics, struggle to capture the human dimension. In his Master’s thesis “Decoding a Developer’s Mind: Multimodal Approaches to Code Comprehension,” Tarek Alakmeh explores human-centered ways to assess code comprehensibility and cognitive load. The study brings together code features, eye tracking, and brainwave data to capture how people actually process code.
By analyzing how eyes move, a neural network can predict perceived code difficulty. Such models that combine code embeddings with eye data outperform state-of-the-art reference models, underscoring the importance of using a multi-modal human-centered approach. Further investigations of brain responses through Electroencephalography (EEG) show additional potential for neural markers of cognitive load. The results point toward automatic indicators of difficulty and more objective ways to measure cognitive load that can inform developer tools and education.