RESEARCH

EXPLORE-QC: Exploiting Phantom Leaks via Optimized RL and KD Exploration in EM Side-Channels on HQC

EXPLORE-QC is a research project that investigates the real-world security of post-quantum cryptographic (PQC) hardware, focusing on how advanced attackers can exploit unintended electromagnetic (EM) emissions. While PQC algorithms are designed to resist quantum computers, their hardware implementations can still leak sensitive information in subtle and unexpected ways.

This project introduces an AI-based, reinforcement learning–driven framework that models a stealthy adversary capable of detecting hidden vulnerabilities using far fewer measurements than traditional security tests. Our results show that hardware designs deemed “secure” by conventional methods may still be vulnerable to intelligent, low-footprint attacks. EXPLORE-QC highlights the need for next-generation certification techniques to ensure the trustworthiness of future cryptographic hardware.

EMPACT: EM and RL based Protection Against Counterfeit ICs

EMPACT develops a scalable, co-existing security ecosystem for detecting counterfeit integrated circuits and stealthy hardware Trojans across modern semiconductor and cloud infrastructures. By unifying electromagnetic (EM) side-channel analysis, reinforcement learning, and real-time on-chip sensing, EMPACT addresses adaptive and low-observable threats that evade conventional testing.

The framework operates as a closed-loop, co-evolutionary ecosystem in which an attack agent generates EM-stealthy Trojans, a defense agent adaptively detects subtle anomalies in real time, and a sensing agent provides continuous on-chip EM monitoring and feedback. Through their continuous co-existence and co-adaptation, EMPACT delivers scalable, hardware-agnostic, and field-deployable solutions for real-time EM-based counterfeit IC and hardware Trojan detection, advancing trusted and resilient hardware infrastructures.

FLARE: Fault Attacks and Hardware Evaluation

FLARE secures embedded and IoT hardware against physical fault attacks, which threaten mission-critical systems. The project focuses on hardware-level trust by studying how fault attacks affect system behavior and cryptographic security in environments where attack and defense agents co-exist and adapt.

The project studies the impact of fault attacks on system behavior and cryptographic security in environments where attack and defense agents co-exist and adapt over time. FLARE aims to identify vulnerable hardware locations, evaluate how fault attacks alter security properties, and develop robust mitigation techniques. A key goal is to train reinforcement learning (RL)–based attack agents capable of bypassing defenses, alongside adaptive defense mechanisms that respond to evolving adversarial strategies.

SEQURE: Secure and Resilient Quantum Hardware

SEQURE secures quantum hardware by addressing vulnerabilities in quantum processors, controllers, and auxiliary components. The project targets fault injections, hardware Trojans, and side-channel leaks, using a unified framework that integrates reinforcement learning, EM and noise-aware detection, and multi-modal side-channel analysis. SEQURE operates as a co-existing ecosystem of attack, defense, and sensing agents, enabling adaptive monitoring, real-time detection, and robust mitigation of quantum hardware threats. By combining fault analysis, Trojan detection, and side-channel monitoring, SEQURE advances trusted and resilient quantum computing for NISQ and cloud platforms.