Security

Security research focuses on identifying vulnerabilities, developing rigorous approaches to evaluating AI systems and advancing trustworthy AI. The Institute explores how intelligent systems can be made more resilient, reliable and secure as AI capabilities become increasingly sophisticated.

A student presenting a research project in a classroom.

Combining LLMs and logic programming for vulnerability discovery

EntailLLM: verifying LLM-generated vulnerability discovery paths with domain knowledge via logic programming
Mukherji, Patil, Payne, Shakarian, Warmsley, Stepp, Kim · arXiv preprint, August 2026

Scalable logic programming for trustworthy AI systems and security applications

Terascale query processing in the browser: rethinking GPU acceleration
Lu, Dyken, Sun, Micinski, Gilray, Kumar · arXiv preprint, July 2026
Scaling worst-case optimal Datalog to GPUs
arXiv preprint, April 2026

AI-driven fuzzing

KnitFuzz: LLM-guided Kernel Fuzzing via Context-Sensitive Socket System Calls
Zhang and Hoque, ACM CODASPY, June 2026.

More information can be found on the people page.