Vahid Behzadan

About

Vahid Behzadan is a leading researcher at the intersection of artificial intelligence and cybersecurity, with a primary focus on the safety and robustness of deep reinforcement learning (DRL) systems. His seminal work, "The Faults in Our Pi Stars: Security Issues and Open Challenges in Deep Reinforcement Learning" (2018, 20 citations), is a foundational survey that systematically mapped the vulnerability landscape of DRL algorithms, exposing critical weaknesses in adversarial environments. This contribution has been instrumental in shaping the emerging field of adversarial machine learning for sequential decision-making, influencing subsequent research on robust policy training and attack mitigation. Behzadan’s research extends to autonomous systems, where he investigates the security implications of AI-driven control in robotics and navigation. His work is notable for bridging theoretical security analysis with practical deployment challenges, earning recognition from both academic and defense communities. By highlighting how even state-of-the-art DRL agents can be manipulated through subtle perturbations, Behzadan has provided a crucial roadmap for developing safer, more resilient AI systems. His ongoing efforts continue to inform best practices for securing intelligent agents in high-stakes environments.

Research Focus

Key Achievements

Top Papers

1. 1
   The Faults in Our Pi Stars: Security Issues and Open Challenges in Deep Reinforcement Learning

   20 citations · 2018

Key Collaborators

Contact & Links