Computer Scientist researches the complexity of artificial intelligence for its usage against cyber-attacks. Meet Yigitcan Kaya

Yigitcan Kaya had a unique influence for his interest in the science, technology, engineering and mathematics (STEM) field, video games. Growing up in Turkey, his family could only afford “cracked” versions of popular video games from merchants. As making these cracked games work often required taking increasingly complex steps, it led Kaya to learn more about how computers operated and the significance of each component, earning him the reputation of “computer guru” from his family.

“Through this pursuit of playing video games, I also developed a deeper interest in computers and started going to the programming club in high school,” Kaya said. “Helping so many people with their computer problems made a career in STEM inevitable.”

Kaya received his undergraduate degree in Computer Science at Bilkent University, followed by his doctoral degree in computer science at the University of Maryland College Park where he discovered Oak Ridge Institute for Science and Education (ORISE).

“In the last year of my PhD, I decided to do a postdoc to pursue academic positions in R1 research institutions in the US and searched for a fellowship opportunity that funds research in cybersecurity and found the Intelligence Community Postdoctoral Research Fellowship Program (IC Postdoc Program),” Kaya said.

The IC Postdoc Program offers scientists and engineers from a wide variety of disciplines unique opportunities to conduct research relevant to the Intelligence Community.

Kaya’s fellowship allows him to work in a team at the cybersecurity lab, also known as the SecLab, at the University of California, Santa Barbara which is led by Giovanni Vigna and Chris Kruegel. The SecLab is highly respected in the cybersecurity community, bringing forth over two decades of experience, groundbreaking research contributions, a robust outreach program and a long list of alumni who are also creating impactful work across the world. A typical work week at his appointment follows a hybrid schedule with office days being the busiest. When working in-person, Kaya brainstorms with his team, collaborates with other student’s projects and works closely alongside his advisors receiving feedback and discussing collaborations.

“As each student is from a distinct background, I also spend quite a lot of time learning about different research areas,” Kaya said. “It often feels like I started a second PhD as every day I’m learning something new in this environment, sometimes just by eavesdropping on other conversations in the lab.”

Kaya is currently working on research which seeks to design techniques that continuously improve and refine existing models as new threats emerge to overcome limitations of artificial intelligence (AI) and machine learning (ML). Traditionally, ML is viewed as a “static tool” that is trained on historical data points and is deployed to prevent future threats. However, due to attackers in cybersecurity rapidly adapting to the existing defenses, novel malware is developed in hopes of circumventing such defenses.

“A critical goal of these techniques is enabling “anti-fragility” in ML-based defenses, which means that they can evolve gracefully through self-updating and external feedback from human operators,” he said. “My emphasis is on incorporating a quantitative notion of trust into ML for explicitly identifying a model's decisions as reliable or unreliable. Reliable decisions ensure resilience, whereas unreliable decisions can be turned into learning opportunities to improve the model toward antifragility.”

Kaya’s research will help move toward the vision and development of long-lasting solutions to cybersecurity defenses. To defend the nation’s critical infrastructure and networks against modern cyber-attacks that have the potential to hinder or halt human activities, powerful defenses that wield the power of recent AI advancements can greatly assist in these efforts. However, as AI continues to evolve, research on its reliability and other under researched areas is crucial to its deployment.

With his fellowship just beginning, Kaya already has many aspirations including becoming a professor at a top research university in the United States and to submit a paper on his ongoing work to the Institute of Electrical and Electronics Engineers (IEEE) Symposium on Security and Privacy in 2025.

“Receiving the IC Postdoctoral Fellowship is one of the first things I achieved as a US Citizen,” Kaya said. “I was born in Turkey and have only been a citizen through naturalization since June 2022 and I’m incredibly proud that right after getting my citizenship, I was awarded a prestigious fellowship to do work that can improve American interests in cyberspace.”

The Intelligence Community Postdoctoral Research Fellowship Program is funded by the Office of the Director of National Intelligence (ODNI) and managed by the Oak Ridge Institute for Science and Education (ORISE) under an agreement between the IC and the U.S. Department of Energy (DOE). ORISE is managed for DOE by ORAU.