Research Summary
(1) online social networks and distributed digital ecosystems, where adversaries exploit human behavior, information structures, and algorithmic weaknesses; and (2) cyber-physical infrastructures, including nuclear instrumentation and control (I&C) and industrial IoT, where cyber compromise can induce real physical consequences.
I develop AI systems that remain robust even under adaptive, coordinated, and AI-driven attacks by integrating machine learning, adversarial modeling, privacy-enhancing technologies, graph intelligence, and cognitive security. As these threats evolve, my work shows that both online manipulation and cyber-physical attacks exhibit stable, mathematically detectable behavioral signatures—from temporal deception patterns and linguistic drift in social systems to anomalies in OT networks—enabling AI models that can predict and mitigate adversarial behavior rather than merely respond to it. Much of my work is grounded in human-centered security, using cognitive modeling, behavioral analytics, and privacy-preserving architectures to protect users and operators in complex and adversarial environments.
My research vision is guided by a central question: How do we design AI systems that preserve security, privacy, and integrity when data, users, and system components are continually evolving and cannot be fully trusted?I direct the Security and Privacy Enhanced Machine Learning (SUPREME) Lab at UTEP, where my group develops secure, privacy-centric AI architectures that unify behavioral modeling, distributed learning, and human-centered defense.
Interests
- Security & Privacy
- Machine Learning
- Online Social Networks
- Distributed & Federated Systems
- Human-Centered Security
- Cyber-Physical & Nuclear Security
- AI for Social Good
