A University of Maryland expert in AI and cybersecurity has received funding from the National Science Foundation (NSF) to make AI tools for writing computer code safer and more reliable.
Yizheng Chen, an assistant professor of computer science with an appointment in the University of Maryland Institute for Advanced Computer Studies (UMIACS), is principal investigator on the NSF Faculty Early Career Development Program (CAREER) award, totaling approximately $573,000 over the next five years.
This highly competitive award—one of NSF’s most prestigious for early-career faculty—recognizes researchers with the potential to serve as academic role models in both research and education, while leading transformative advances in their fields.
Chen’s research focuses on code large language models (Code LLMs), which are AI systems that help programmers by completing partial programs or generating code from natural language instructions. These models power popular tools such as ChatGPT, GitHub Copilot, and Cursor, which are already integrated into software development environments.
She notes that coding assistants have revolutionized software development by boosting productivity, and by 2028 it is estimated that 75 percent of enterprise software engineers will rely on these tools. However, Code LLMs can also produce code with vulnerabilities, raising significant security concerns.
To address these challenges, Chen’s project is developing the next generation of secure Code LLMs that can generate safe, correct code in realistic programming scenarios.
The project addresses three core objectives. First, it develops benchmarks and metrics to evaluate Code LLMs in a range of coding scenarios, from basic programming exercises to complex real-world projects. Next, it creates methods that help AI systems generate code that is both secure and high-quality. Finally, it explores techniques that teach Code LLMs safe coding practices without limiting their overall capabilities.
Chen, who is also a core member of the Maryland Cybersecurity Center, plans to share datasets, tools and findings with the broader research community.
In addition to advancing research, the project provides hands-on learning opportunities for graduate and undergraduate students, who will contribute to the development and evaluation of secure Code LLMs.
***
“Secure Code Generation with Large Language Models” is supported by NSF grant #2442719 from the NSF’s Division of Computer and Network Systems
PI: Yizheng Chen, an assistant professor of computer science with an appointment in the University of Maryland Institute for Advanced Computer Studies.
About the CAREER award: The Faculty Early Career Development (CAREER) Program is an NSF activity that offers the foundation’s most prestigious awards in support of junior faculty who exemplify the role of teacher-scholars through outstanding research, excellent education and the integration of education and research within the context of the mission of their organization.